469

Biochimica et Biophysica Acta, 4 5 4 ( 1 9 7 6 ) 4 6 9 - - 4 7 9 © Elsevier/North-Holland Biomedical Press

BBA 98767

CHARACTERIZATION OF THE RNA TRANSCRIBED IN VITRO FROM NATIVE MAMMALIAN DNA BY ESCHERICHIA COLI RNA POLYMERASE

A N G E L A L O N S O *, G.D. B I R N I E , L. K L E I M A N **, A.J. M A C G I L L I V R A Y a n d J O H N PAUL

Beatson Institute for Cancer Research, 132 Hill Street, Glasgow, G3 6UD (U.K.) ( R e c e i v e d April 2 0 t h , 1 9 7 6 )

Summary High-molecular-weight native mouse DNA was transcribed with Escherichia coli RNA polymerase under low salt conditions, and the nature of the DNA sequences transcribed determined by molecular hybridization. The results indicated that E. coil RNA polymerase does not transcribe the sequences in native mouse DNA randomly under these conditions. First, hybridization with a large excess of mouse DNA showed that no more than 5% of the RNA synthesized had been transcribed from repeated sequences in the DNA. Second, hybridization with tracer amounts of labelled non-repeated mouse DNA indicated that the bulk of the R N A h a d been transcribed from less than 1% of the non-repeated sequences and only about 10% had been transcribed from a further 25% of these sequences; the remaining non-repeated sequences in the DNA, amounting to 50% of the genome, were not represented in the RNA synthesized in vitro to any detectable extent. Third, the proportion (40%) of complementary DNA transcribed from mouse-liver nuclear polyadenylated RNA which hybridized with the RNA synthesized in vitro was significantly greater than would have been expected if transcription had been random. The data have also been interpreted as indicating the presence of two types of initiation site for E. coli RNA polymerase in the non-repeated sequences in mouse DNA. The frequencies of their occurrence have been calculated to be one per 150 000 base-pairs and one per 5000 base-pairs, respectively.

* Present address: Institut fiir experimentelle Pathologie, Deutsches Krebsforschungszentrum, 69 Heidelberg (G.F.R.). ** Present address: Lady Davis Institute for Medical Research, Jewish General Hospital, Montreal, Quebec H3T 1E2 (Canada).

470 Introduction Bacterial RNA polymerases, in particular Escherichia coli R N A polymerase, have been widely used to transcribe chromatin in vitro, in experiments designed to determine the way in which chromatin proteins impose specific limitations on the sequences transcribed in the DNA to which they are bound. The transcription of native DNA in vitro by E. coli RNA polymerase has been studied extensively (see ref. 1), but most of the information available has been obtained using bacteriophage DNAs as templates. Despite the extensive use of this enzyme for the transcription of chromatin in vitro, there have been few studies of the actual sequences in the RNA transcribed by E. coli RNA polymerase from native mammalian DNAs. Although Cedar and Felsenfeld [2] showed that there are a large number of sites (about one per 1200 base-pairs) in native calfthymus DNA on which initiation of transcription by E. coli polymerase occurs, it was not possible to tell from their experiments whether the initiation sites are clustered or are distributed evenly throughout both repeated and non-repeated sequences in the DNA. However, data obtained from DNA • RNA hybridization experiments by Bishop and his collaborators [3,4] have indicated that the initiation sites recognised by E. coli RNA polymerase are not equally distributed in mammalian DNA. These experiments showed that, although a large proportion of the repeated DNA sequences were transcribed by the enzyme, only a small proportion of the R N A synthesized had been transcribed from these sequences, and a much larger proportion comprised transcripts from non-repeated sequences. As part of our studies on the transcription of chromatin in vitro, we undert o o k a detailed examination of the sequences in the RNA transcribed in vitro by E. coli R N A polymerase from native mouse DNA, using D N A . R N A hybridization techniques which have recently been developed to determine the relative abundances of the RNAs transcribed from different DNA sequences. The data in this report show that, in the transcribed RNA, (i) the repeated DNA sequences are under-represented; (ii), at least 75% of the non-repeated sequences (that is, half of the genome) is represented to a vanishingly small extent, if at all; and (iii), a very small proportion (less than 1%) of the non-repeated sequences are grossly over-represented. Materials and Methods All possible precautions were taken to eliminate ribonucleases and heavy metal ions [5]. All glassware used in the synthesis of RNA, and in the isolation and hybridization of RNA and DNA (including capillaries), had been siliconized ('Repelcote'; Hopkins and Williams Ltd., Chadwell Heath, Essex, U.K.), then sterilized by rinsing with 0.1% aqueous diethylpyrocarbonate and drying at 100°C. Solutions in which hybridization reactions were done had been passed through Chelex-100 resin (Bio-Rad Laboratories, Richmond, Calif., U.S.A.) to remove heavy metal ions, then sterilized by treatment with diethylpyrocarbonate; excess diethylpyrocarbonate was destroyed by heating at 60~C for 18 h. Suspensions of Sephadex (Pharmacia, Uppsala, Sweden) were sterilized by shaking them with diethylpyrocarbonate (0.1%), then heating at 60 ° C for 18 h.

471 High-molecular-weight DNA. Nuclei were prepared from livers of y o u n g adult mice by the sucrose/citric acid procedure [6] as described previously [5]. Native DNA was extracted from these nuclei b y the m e t h o d of Gross-Bellard et al. [7]. The molecular weights of the DNA molecules were calculated from their contour lengths, measured in the electron microscope by Dr. Lesley Coggins using the aqueous technique of Davis et al. [8]. A total of 80 molecules, selected at random, were measured. The internal standard was SV40 DNA, kindly given by Dr. R. Eason. The DNA was stored at --20°C in 0.1 M NaCl. R N A polymerase. The holoenzyme was extracted from E. coli (M.R.E. 600) as described by Burgess [9], omitting the phosphocellulose chromatography step. The enzyme was then absorbed to, and eluted from, a column of DNAcellulose as described by Bautz and Dunn [10] except that high-molecularweight calf-thymus DNA (prepared as described for mouse-liver DNA) was used in place of T4 DNA. This procedure successfully removed ribonuclease from the RNA polymerase, as judged by the failure of the enzyme preparation to degrade E. coli ribosomal RNA. Transcription o f DNA. The incubation mixture contained 40 mM Tris • HC1, pH 7.9, 10 mM NaC1, 1 mM dithiothreitol, 0.1 mM EDTA, 12 mM MgC12, 1 mM MnC12, 0.4 mM ATP, 0.4 mM GTP, 0.4 mM CTP, 0.4 mM UTP and, in each 1 ml, 100 pg of mouse-liver DNA and 160 units [9] of E. coli R N A polymerase. When labelled R N A was required, the UTP was replaced by 40 pM [ 3H]UTP (300 pCi/ml; Radiochemical Centre, Amersham, U.K.). After incubation at 37°C for 30 min, RNA was isolated from the mixture by the procedure described previously for the isolation of R N A from nuclei [5]. The R N A eluted in the void volume of the Sephadex G-100 column was precipitated by the addition of NaC1 to 0.2 M and 2 vol. of ethanol; after storage overnight at --20°C, the precipitated RNA was recovered by centrifugation (10 000 × g for 15 min), washed with ethanol and dried under a stream of nitrogen. The RNA was desalted by passage through a column of Sephadex G-50 and a 5 mm pad of Chelex-100 equilibrated with sterile distilled water, and lyophilized. The molecular weight of the RNA was measured by electrophoresis in 3.8% polyacrylamide gels in 100% formamide, essentially as described by Staynov et al. [11]. Hybridization o f mouse non-repeated DNA with RNA. Highly labelled nonrepeated ("unique") DNA was isolated from mouse Friend cells (clone M2) grown at 37°C in a 250 ml spinner culture [12]. When the cells were in mid-log phase of growth (8 • l 0 s cells/ml), 2.5 ml of a mixture of aminopterin (2 • 10 -s M), deoxyadenosine (7 • 10 -3 M) and glycine (3 • 10 -2 M) was added, followed 15 min later by 2.5 mCi of [Me-3H]thymidine (40.3 Ci/mmol; Radiochemical Centre, Amersham, U.K.). The cells were collected by centrifugation 23 h later, and washed with phosphate-buffered saline. DNA was isolated as described by Hell et al. [13], sheared b y ultrasonication, desalted on Sephadex G-50 (with a pad of Chelex-100) in sterile distilled water, and lyophilized. The DNA was redissolved in 0.5 M NaC1, 25 mM HEPES, 0.5 mM EDTA, 50% formamide, pH 6.8, at 5 mg/ml, sealed in capillaries, heated at 70°C for 5 min and incubated at 43°C for 4.25 h (Cot -'- 250 mol • s • litre-'). Single- and double-stranded DNA were separated by hydroxyapatite column chromatography [14]; the single stranded DNA was isolated and re-cycled through the hybridization and fractionation procedure. The unhybridized fraction of the DNA was finally desalted

472 on Sephadex G-25 and lyophilized. This procedure yielded non-repeated mouse DNA of mean single-stranded molecular weight 5 • 104 daltons (as determined by alkaline sucrose-gradient centrifugation [13] ), and specific activity 8.7 • 105 counts/min/pg. When annealed with a 2000-fold excess of total, sheared mouse DNA, 85--90% of the labelled non-repeated DNA formed duplexes. Appropriate volumes of R N A and labelled non-repeated DNA in sterile, distilled water were mixed, lyophilized and redissolved in 0.24 M phosphate buffer (equimolar Na2HPO4 and NaH2PO4; pH 6.8) containing 0.1% sodium lauryl sulphate. Portions of the solution were sealed in glass capillaries, denatured by heating at 100°C for 5 min then incubated at 60°C. The capillaries were flushed o u t with 1 ml of 0.03 M phosphate buffer, pH 6.8, and single- and doublestranded molecules were separated by chromatography on hydroxyapatite at 60°C [14]; the radioactivity in each fraction was measured by liquid scintillation spectrometry in Instagel (Packard Instrument Ltd., Caversham, Berks., U.K.). Hybridization of [3H]RNA with whole mouse DNA. DNA was isolated from whole mouse embryos as described by Hell et al. [13], sheared by ultrasonication and desalted by chromatography on Sephadex G-50 (with a pad of Chelex100) in sterile distilled water. The mean single-stranded molecular weight of the DNA was 1 . 5 . 1 0 s. Appropriate volumes of [ 3 H ] R N A and sheared DNA in sterile distilled water were mixed, lyophilized and redissolved in 0.12 M phosphate buffer (equimolar NaH2PO4 and Na2HPO4; pH 6.8) containing 0.1% sodium lauryl sulphate. Portions of the solution were sealed in glass capillaries, denatured by heating at 100°C for 5 min then incubated at 65°C. The capillaries were flushed o u t with 1 ml of 0.24 M phosphate buffer, and the amount of [ 3 H ] R N A which had formed stable hybrids was determined by measuring the proportion which was not degraded by incubation with ribonuclease A (200 pg/ml) at 37°C for 20 min [4]. Hybridization of cDNA with RNA. Mouse-liver nuclei were prepared by the sucrose-citric acid method [5,6] and total RNA was isolated essentially as described previously [5] except that the nuclei were first lysed in sodium lauryl sarcinosate and incubated for 45 min with 500 pg/ml proteinase K [15] (British Drug Houses, Poole, Dorset, U.K.). Polyadenylated [poly(A) ÷] R N A was isolated following affinity chromatography on oligo(dT)-cellulose (Collaborative Research Inc., Waltham, Mass., U.S.A.) as described previously for polysomal poly(A) ÷ R N A [16]. The RNA was transcribed, and the cDNA (spec. act., 5000 cpm/ng) fractionated and isolated, as described previously [5]. The mean molecular weight of the cDNA was 1 • 10 s daltons. The hybridization reactions were done at 43°C in 0.5 M NaC1, 25 mM HEPES, 0.5 mM EDTA, 50% formamide, pH 6.8 [5,16], and the proportion of cDNA in hybrid was determined [ 5] by measuring the proportion of radioactivity rendered acid soluble by incubation with ~S, nuclease (Sigma Chemical Co. Ltd., London, U.K.). Results

Transcription of DNA The DNA used was prepared by the m e t h o d of Gross-Bellard et al. [7] in order that the template should be of high molecular-weight and as free as possible

473 of single-strand nicks. As calculated from contour lengths in electron micrographs, 90% of the DNA was in molecules greater than 6.5 • 106 daltons while the average molecular weight of the DNA was 13.2 • 106; estimation of the single-strand molecular weight by rate-zonal sedimentation in alkaline sucrose gradients indicated there was 1--2 scissions per DNA strand on average. The E. coli R N A polymerase was prepared as described b y Burgess [9], omitting the phosphocellulose column step in order to retain sigma factor. The enzyme was devoid of ribonuclease activity and was template-dependent; transcription of the DNA in vitro yielded 40--45 pg of R N A per 100/~g of DNA. Electrophoresis on denaturing formamide-polyacrylamide gels showed that most (80%) of the RNA synthesized consisted of molecules around 4 • l 0 s daltons; the remainder had a molecular weight of a b o u t 6 • 104 (Fig. 1).

Hybridization o f R N A to non-repeated DNA To determine what proportion of the non-repeated sequences in the highmolecular-weight DNA was represented in the in vitro transcripts, a large (10 000-fold) excess of the E N A was hybridized with highly radioactive nonrepeated mouse DNA. The reaction appeared to be complete when 12--13% of the non-repeated DNA had hybridized with the R N A (Fig. 2), indicating that, if transcription was asymmetric, about 25% of the non-repeated sequences in the genome had been transcribed by E. coli polymerase. The possibility that a higher proportion of the non-repeated sequences had been transcribed cannot be ruled out, b u t it is clear that, if they were, the abundance of their transcripts was extremely low (less than 1% of the total RNA). A b o u t 70% of the sequences in mouse DNA are non-repeated sequences [17].

oi..

'

I

'

I

I

'

;

I

'

'

I

'

'

•

5S

18S

J,

J,

28S

z~

J,

i8

'

< m 0 en
I Z

40

< Z a u

60

w U 8() ne W I1. 10C

L

I

I 10

~

i

I

,

10 2

I

] 10 3

~

I

I 104

J

I 10 5

Rot Fig. 4. T i m e - c o u r s e o f h y b r i d i z a t i o n o f c D N A t r a n s c r i b e d f r o m m o u s e - l i v e r n u c l e a r p o l y ( A ) * R N A w i t h (a) u n f r a c t i o n a t e d m o u s e - l i v e r n u c l e a r R N A ( I l ) ; (b) mouse-liver nuclear poly(A)* RNA ( o - o); and (c) RNA transcribed in vitro from native mouse DNA ( A - A). T h e c o n c e n t r a t i o n o f R N A w a s 1 t o 1 0 m g / m l a n d t h e r a t i o o f R N A : c D N A , 3 0 0 t o 9 0 0 0 : 1. S e l f - a n n e a l i n g o f t h e c D N A ( . . . . . . ) was m e a s u r e d i n h y b r i d i z a t i o n r e a c t i o n s d o n e i n p a r a l l e l w i t h E. coli R N A a t t h e s a m e c o n c e n t r a t i o n s .

sors of polysomal m R N A sequences (see ref. 23). Thus, measurement of the extent to which this c D N A forms hybrids with R N A transcribed from D N A in vitro gives an indication of the e x t e n t of h o m o l o g y between the sequences in in vitro RNA transcripts and nuclear p o l y ( A ) ÷ RNA. When the c D N A was annealed with its template RNA, 80% formed S~-nuclease-resistant hybrids; a similar proportion hybridized with total nuclear RNA. In contrast, 30% of the c D N A hybridized when annealed with a 9000-fold excess of the RNA synthesized in vitro (Fig. 4). After normalising to allow for the proportion of the c D N A which was non-hybridizable, the data in Fig. 4 indicate that at least 40% of the c D N A was complementary to the sequences transcribed in vitro from mouse DNA. Fig. 4 also shows that the rate of the reaction between the c D N A and the in vitro RNA transcripts was markedly slower than that of the reaction between the c D N A and its template RNA. Since it is n o t k n o w n whether the relative abundances of the hybridizing R N A sequences in the nuclear p o l y ( A ) ÷ R N A were similar to those in the R N A synthesized in vitro, it is n o t possible to make a definitive assessment of the concentration in the latter of those sequences which hybridized to the cDNA. However, if the relative abundances were similar, by comparing the Rott/2 values of the two reactions it can be estimated that they constituted less than 10% of the total R N A transcribed from D N A in vitro. Discussion The picture which emerges from these analyses is one of marked non-randomness of transcription of high-molecular-weight native mouse D N A by E. coli R N A polymerase h o l o e n z y m e . First, although 20% of the sequences in mouse D N A belong to the middle-repetitive class [ 1 7 ] , only a very small proportion (less than 5%) of the RNA had been transcribed from these sequences. This con-

477 clusion is n o t at variance with previous data which indicated that a high proportion of the middle-repetitive sequences of mammalian DNA are transcribed by bacterial polymerase [3,24]. In those experiments, saturation of middle-repetitive DNA sequences was obtained in DNA • RNA hybridization reactions only at high ratios of RNA to DNA; moreover, less than 5% of the input RNA hybridized under the conditions used [3]. Second, although almost all of the RNA had been transcribed from the non-repeated DNA sequences of the mouse genome, no more than a quarter of these DNA sequences were represented in the RNA. Moreover, even those non-repeated DNA sequences which had been transcribed were n o t all equally represented in the RNA; the major part of the RNA was composed of transcripts from less than 1% (possibly less than 0.1%) of the non-repeated DNA sequences, while the transcripts from the remaining sequences made up no more than 10% of the total RNA synthesized. However, these data should not be taken to imply the existence of completely sharp demarcations between the abundance classes in the RNA synthesized. A more likely pattern is depicted in Fig. 5, which suggests the presence of an overlapping series of abundance classes which fall into 3 main groups. These data show that some of the sequences in non-repeated native DNA are preferentially transcribed by E. coli RNA polymerase in vitro. While it is possible that the enzyme is capable of transcribing some sequences in native DNA more rapidly or frequently than others, a likelier proposition is that initiation of transcription occurs more readily on some sequences than on others. If this is correct, the pattern of transcription implies that there are two types of initiation site on the DNA, which could be analogous to (but n o t necessarily the same as) the class A and class B sites detected on phage DNA [25]. I t is unlikely that either one or both of these sites are ends and single-strand scissions in the DNA template since the DNA was prepared by a m e t h o d which produced high-molecular-weight DNA; moreover, ends and single-strand nicks might be expected to be randomly distributed throughout both the repeated T

100!-

1

[

T

mso;

a iJi tn E too60 z 11: ~ 40 0 i-z 20 u/ u n~ 0

0.1 0.2 PROPORTION

0"3 0'4 O F THE MOUSE

GENOME

F i g . 5 . D i a g r a m m a t i c r e p r e s e n t a t i o n o f the d i s t r i b u t i o n o f R N A s e q u e n c e s transcribed in v i t r o f r o m native mouse DNA; HR ~ highly repeated, M R ~ m i d d l e r e p e a t e d and N R ~ n o n - r e p e a t e d s e q u e n c e s in D N A .

478 and non-repeated sequences in the DNA. In addition, there is evidence t hat single-strand nicks in native DNA are n o t effective initiation sites for the synthesis o f RNA chains [26]. Although the experiments were n o t specifically designed for the purpose, by making a n u m b e r of assumptions an estimate of the n u m b e r of initiation sites on native mouse DNA can be obtained from the data. The sites giving rise to the major portion of the RNA initiated transcription from no m ore than 1% of the non-repeated DNA sequences. Assuming a s y m m e t r y of transcription, the base-sequence c o m p l e x i t y of the transcribed sequences was thus, at most, 0.01 X 1.9 • 109 base-pairs, t hat is 1.9 • 107 base-pairs. The molecular weight of the bulk of the RNA synthesized in vitro was 4 • l 0 s ; assuming t hat there was no degradation of the RNA after transcription, the base-sequence c o m p l e x i t y of each DNA sequence transcribed was, therefore, 1.2 • 103 base-pairs. If there was only one initiation site per transcribed sequence then the num ber of initiation sites o f this t y p e was 1.9 • 107 i.e. ab o u t 1.6 • 104 per haploid genome. 1.2 103' Making the same assumptions, the n u m b e r of sites of the second type, which initiated transcription f r om 25% of the non-repeated DNA, can similarly be calculated to be a b o u t 4 • l 0 s per haploid genome. These data t herefore suggest there is a b o u t one initiation site per 5000 base-pairs of non-repeated native mouse DNA whereas a more direct m e t h o d of estimating this n u m b e r indicated that there is one initiation site for E. coli polymerase per 1200 base-pairs on calf-thymus DNA [2]. The 4-fold discrepancy could reflect an intrinsic difference between mouse and calf-thymus DNA; alternatively, it could be due to inaccuracies in the estimate from the data in this paper due, for example, to some clustering of the initiation sites. However, it should be n o t e d t hat the technique used by Cedar and Felsenfeld [2] would n o t have distinguished between the two classes o f initiation site which these data suggest are present. Although Meilhac and Chambon [27] concluded t hat calf-thymus B and E. coli h o l o e n z y m e RNA polymerases initiate transcription at different sites on calf-thymus DNA, it is interesting to speculate on w het her the sequences in native mouse DNA which are transcribed in vitro by E. coli polymerase bear any relationship to the sequences which are transcribed in vivo. One possible clue is given by the e x p e r i m e n t in which cDNA homologous to nuclear poly(A) ÷ RNA was annealed with the RNA synthesized in vitro. If in vitro transcription were purely r an d o m with respect to those sequences represented in the cDNA, at most 25% of the cDNA would have been c o m p l e m e n t a r y to the sequences transcribed from DNA in vitro whereas, in poi nt of fact, 40% of the cDNA form ed hybrids. This difference is small, but significant. Moreover, the cDNA represents only the 3'-ends of the nuclear poly(A) ÷ RNA whereas RNA is transcribed in vivo f r o m its 5'-end [28]. Thus, if transcription in vitro were initiated close to the same points as transcription in vivo, the e x t e n t to which the cDNA hybridized with RNA synthesized in vitro will have underestimated the degree of correspondence between the DNA sequences transcribed in vivo and those transcribed in vitro by E. coli RNA polymerase.

479

Acknowledgements The Beatson Institute is supported by grants from M.R.C. and C.R.C.A.A. was on leave of absence from Deutscheskrebsforschungzentrum, Heidelberg, W. Germany, and was supported by an EMBO Fellowship; L.K. was an American Leukemia Society Fellow. We are grateful to Dr. B.D. Young for invaluable advice and criticism, to Dr. Lesley Coggins for D N A molecular-weight determinations by electron microscopy, and to Dr. Anna Hell for a gift of high specific activity mouse DNA. References 1 Chamberlin, M.J. (1974) in The Enzymes (Boyer, P., ed.), Vol. I 0 , pp. 333--374, Academic Press, N.Y. and London 2 Cedar, M. and Felsenfeld, G. (1973) J. MoL Biol. 77, 237--254 3 Melli, M. and Bishop, J.O. (1970) Biochem. J. 120, 225--235 4 Melli, M, Whitfield, C., Rao, K.V., Richardson, M. and Bishop, J.O. (1971) Nat. New Biol. 231, 8--12 5 Getz, M.J., Birnie, G.D., Young, B.D., MacPhail, E. and Paul, J. (1975) Cell 4, 121--129 6 Busch, H. and Smetana, J. (1970) The Nucleolus, Academic Press, London and New York 7 Gross-BeUard, M., Oudet, P. and Chambon, P. (1973) Eur. J. Biochem. 36, 32--38 8 Davis, R.W., Simon, M. and Davidson, N. (1971) in Methods in Enz ymol ogy (Grossman, L. and Moldave, K., eds.), Vol. 21, pp. 413--428, Academic Press, London and New York 9 Burgess, R.R. (1969) J. Biol. Chem. 244, 6160--6167 10 Bautz, E.K.F. and Dunn, J.J. ( 1 9 7 1 ) i n Procedures in Nucleic Acid Research (Cantoni, G.L. and Davies, R., eds.) Vol. 2 pp. 743--747, Harper and Row, New York and London 11 Staynov, D.Z., Pinder, J.C. and Gratzer, W.B. (1972) Nat. New Biol. 235, 108--110 12 Gilmour, R.S., Harrison, P.R., Windass, J.D., Affara, N.A. and Paul, J. (1974) Cell Different, 3, 9 ~ 2 2 13 Hell, A., Birnie, G.D., Slimming, T.K. and Paul, J. (1972) Analyt. Biochem. 4 8 , 3 6 9 - - 3 7 7 14 Harrison, P.R., Birnie, G.D., Hell, A.0 Humphries, S., Young, B.D. and Paul, J. (1974) J. Mol. Biol. 84, 539--554 15 Glisin, V., Crkvenjakov, R. and Byus, C. (1974) Biochemistry 13, 2633--2637 16 Birnie, G.D., MacPhail, E., Young, B.D., Getz, M.J. and Paul, J. (1974) Cell Different. 3, 221--232 17 Britten, R.J. and Kohne, D.E. (1968) Science 161, 629--540 18 Sober, H.A. (1968) Ed., Handbook of Biochemistry, p. H58, The Chemical R ubbe r Co., Cleveland, Ohio 19 Young, B.D. and Paul, J. (1973) Biochem. J. 135, 573--576 20 Hell, A., Young, B.D. and Birnie, G.D. (1976) Biochim. Biophys. Acta 442, 37--49 21 Williamson, R., Morrison, M., Lanyon, G., Eason, R. and Paul, J. (1971) Biochemistry 10, 3014--3021 22 Bishop, J.O. (1972) Biochem. J. 126, 171--185. 23 Lewin, B. (1975) Cell 4, 77--93 24 Paul, J., Carroll, D., Gllmour, R.S., More, I.A.R., Threlfall, G., Wilkie, M. and Wilson, S. (1972) in Gene Transcription in Reproductive Tissue (Diczfalusy, E., ed.), pp. 277--297, Karolinska Institutet, Stockholm 25 Hinkle, D.C. and Chamberiin, M.J. (1972) J. Mol. Biol. 70, 157--185 26 Hinkle, D.C., Ring, J. and Chamber]in, M.J. (1972) J. Mol. Biol. 70, 197--207 27 Meilhac, M. and Chambon, P. (1973) Eur. J. Biochem. 3 5 , 4 5 4 - - 4 6 3 28 Hadjiolov, A.A. (1967) in Prog. Nucleic Acid. Res. Mol Biol. (Davidson, J.N. and Cohn, W.E., eds.), Vol. 7, pp. 195--242, Academic Press, New York and London