Sequence studies of the 5 S DNA of Xenopus laevis

J. Mol. Biol. (1974) 89, 703-718

Sequence Studies of the 5 S DNA of Xenopus Zaevis G. G, BROWNLEE,

E. M. CARTWRICHT

Medical Research Council Laboratory qf Molecular Biology Hills Road, Cambridge CB2 2&H, England AXD DONALD

D. BROWN

Department of Embryology Carnegie Institute of Washington Baltimore, Md 21210, U.S.A. (Received 5 June 1974) complementary RNA transcripts of Nucleotide sequence studies of 32P-labelled the 5 S DNA of Xenopus laevis have shown that the gene region codes for the oocyte-type 5 S RNA. Genes for the somatic-type 5 S RNA, known to have 6 of its 120 bases differing from the oocytes 5 S RNA, do not comprise more than 3% of the genes within this 5 S DNA. Twelve closely related oligomers have been identified in varying yields as sequences from the sense strand of the spacer region of 5 S DNA. One of the most common of these sequences (present in more than 3 copies per repeat) is 5’-A,-C-T-C-A,-C-Ts-G-3’; this oligomer itself may have been derived from an ancient duplication. The tandem arrangement of these oligomers has been shown. They comprise at least half and perhaps all of the A + T-rich part of the spacer (which is 60% of each repeat).

1. Introduction The 5 S ribosomal DNA of Xenopus laevis is of particular interest for detailed sequence studies because it contains a gene of known function, the structural gene for 5 S ribosomal RNA, and because it can be purified from bulk X. luevis DNA by’repeated caesium chloride centrifugations (Brown & Weber, 1968; Brown et al., 1971). The DNA is highly repetitious, there being about 24,000 copies per haploid complement of DNA. The repeats of 5 S DNA vary slightly in length (Carroll &I D. D. Brown, unpublished data) and contain an average of about 750 base pairs. Each repeat is divided into two adjacent regions of vastly different base composition; about 60% of each repeat is high in A + T content while the remaining 40% is relatively high in G + C content. The differences between these two regions and their regular alternation has been demonstrated by thermal denaturation of pure 5 S DNA and its denaturation map as viewed by eIectron microscopy (Brown et d., 1971). Each repeat contains a single 5 S gene of about 120 base pairs presumably located within and comprising from one-half to one-third of the high G + C region. Therefore, about 85% of each repeat including all of the A + T and most of the G + C region is termed spacer DNA (see Fig. 3). Bases are distributed asymmetrically between the two strands of 5 8 DNA 703

704

ET

G. G. BROWNLEE

AL.

which can be separated in alkaline &Cl. 5 8 RNA hybridizes only to the lighter (L) strand which is thus the sense strand. The sequence of the gene product, 5 S RNA, has been determined (Wegnez et al., 1972; Brownlee et al., 1972; l?ord & Southern, 1972) by comparison with the sequence of human 5 S RNA (E’orget & Weissman, 1967). Two distinct 5 S RNA species are present, one synthesized in oocytes, and the other present in somatic tissues. The oocyte type 5 S RNA differs in six bases from the somatic type and also shows microheterogeneity. In this paper we report experiments which demonstrate that isolated 5 S DNA codes mainly, if not exclusively, for oocyte 5 S RNA. In addition, part of the spacer region has been sequenced and found to consist of a family of related sequences of about 15 nucleotides in length. For analysis of both gene and spacer regions the 5 S DNA has been transcribed with the RNA polymerase of Escherichia coli in the presence of each u-32P-labelled nucleoside triphosphate in turn. The labelled complementary RNA transcript was then sequenced by standard methods (Brownlee, 1972).

2. Materials and Methods (a) Pur$cation

of 5 S DNA

The DNA used for 5 S DNA purification was isolated from X. la&s erythrooytes (Brown & Weber, 1968). The 5 S DNA comprises about 0.7% of the bulk DNA and was purified with about 10% yield by 4 density gradient centrifugation steps (Brown et al., 1971). Purified fragments of 5 S DNA out with the restriction enzyme from Haemophilw in$uenme (Hind III) were prepared by D. Carroll and purified by slab gel eleotrophoresis. (b) Preparation

of “ZP-labelled

RNA transcrirpt

Complementary RNA was prepared from native 5 S DNA, or from its separated strands, with E. co% DNA-dependent RNA polymerase according to Burgess (1969), in the presence of one cc-32P-labelled nuoleoside triphosphate. In some experiments, the 5 S DNA was denatured just before transcription by heating to 100°C in distilled water. The conditions for incubation were as follows (Blackburn, E., manuscript in preparation) in a reaction of 32P-labelled ribonuoleoside triphosphate (about 50 &i, 20 to O.lml:lto2p.g5SDNA; 100 Ci/mmol, from New England Nuclear) and unlabelled ribonuoleoside triphosphates, 0.01 pmol (except for ATP, 0.1 pmol). The following salts were added to give the final concentrations shown: TriseHCI, pH 7.8 (40 mu), MgCIz (10 KnM), KC1 (150 mu), dithioEDTA (0.5 mM) and sodium phosphate, pH 7.8 (0.4 mM). Incubation threitol (2 InM), was usually for 3 h at 37°C. The extent of polymerization was assayed by precipitation with 5% triohloroaoetio acid according to Burgess (1969). complementary RNA for sequence studies involved The purification of the 32P-labelled transcript from low molecular hydrolysis of 5 S DNA, separation of the 32P-labelled weight components in the reaction mixture and deproteinization. About 5 pg of pancreatic DNAase (Worthington, eleotrophoretioally purified) and 50 pg of unlabelled carrier tRNA of E. coli (B.D.H.) were added and incubation continued for a further 15 min at 37°C. The reaction mixture was diluted to 1 ml with 0.5% sodium dodecyl sulphate and to a 20 om column of Sephadex G50 (medium) in then applied, at room temperature, O-1 M-Nacl, 0.01 M-Tris*HCl (pH 75), 0.005 M-EDTA, and 0.5% sodium dodecyl sulphate. The material eluting in the void volume (high molecular weight) was deproteinized by shaking vigorously for 15 min with an equal volume of phenol. After separation of the aqueous and phenol layers by oentrifugation, and re-extraction of the phenol layer with a further half volume of distilled water, the combined aqueous layers were precipitated by the addition of 2.5 vol. of 95% ethanol. RNA was collected by oentrifugation and purified from small amounts of sodium dodecyl sulphate by redissolving it in 1 ml of 2% sodium acetate, pH 5.0, and reprecipitating with 2.5 vol. of 95% ethanol. The yield of

SEQUENCE

OP 6 S DNA

705

RNA was usually 5 to 20 &i in experiments in which CTP or UTP were used as input labelled triphosphates. Much lower yields of 05 to 2 &i were obtained with ATP or GTP. Native 5 S DNA rather than purified H or L strands was used for these experiments, as E. coli RNA polymeraae selectively copies the H strand of native DNA several times more efficiently than L strand (Reeder, 1973). Thus, native DNA was a more readily available and convenient template for these experiments than purified H strands. Identical fingerprints were obtained, such as those shown in Plate II, whether purified H strand or native 5 S DNA was transcribed. It should be noted that the transcripts obtained are copies of the nonaense (H) strand; it follows therefore that the RNA sequences reported in this paper are sequences of the sense or L strand of the DNA if uridine (U) residues are replaced by thymidine (T). (c) Sequence techniques Most of the standard methods for RNA sequencing Brownlee, 1972). Some of the more critical details below. (i) Partial

digestion

of RNA

with

have been published of enzyme digestion

in detail (see are described

T1 RNAase

Suitable conditions, giving the separation shown on Plate III, were obtained by digestion for 45 min at 0°C with one part by weight of T, RNAase (Sankyo, Japan) to 400 parts by weight of RNA in 0.01 m-Tris.HCl (pH 7.5), 1 mM-MgCl,. Typically 40 pg of carrier tRNA and 0.1 pg of T, RNAase were used. The reaction volume of 5 ~1 was applied directly onto cellulose acetate (Schleicher and Sohull) for fractionation in the first dimension of the standard two-dimensional system. Fractionation in the second dimension was by homochromatography using a 4% unhydrolysed, dialysed homomixture on a 1: 7.5 DEAE-cellulose thin-layer plate (Brownlee & Sanger, 1969). The partial T, products were eluted, divided into two and analysed both by complete digestion with T1 RNAase and with pancreatic RNAase. (a) The further digest with T, RNAaae (O-1 mg/ml for 1 h at 37°C in 0.01 M-Tris.HCl (pH 7.5), 1 mrvr-EDTA) was analysed by one-dimensional homochromatography on thin-layer plates as described above. Oligonucleotides were identified by position and by the end products when they were digested further with pancreatic RNAase and analysed by DEAE-paper ionophoresis at pH 3.5. (b) The further digest with pancreatic RNAase (0.1 mg/ml for 1 h at 37°C in 0.01 M-Tris.HCl (pH 7.5). 0.001 M-EDTA) was analysed by ionophoresis on DEAE-paper using 7% formic acid. The pancreatic end products were identified by position and by analysis of their digestion products with T, RNAase after separation by ionophoreais on DEAE-paper at pH 3.5. (ii) Xequence of Tl end products Partial digestion with U, RNAase of oligonucleotides eluted from thin-layer plates was carried out in 10 ~1 with 5 units of U, RNAase (Sankyo, Japan) per ml in 0.05 Msodium acetate (pH 4.5), 0.002 M-EDTA for 6 h at 37°C. Products were fractionated by ionophoresis on DEAE-paper at pH 1.9 and identified by position and alkaline hydrolysis. Complete digestion with U, RNAase required digestion with 10 units enzyme/ml for I6 h at 37’C. Partial pancreatic RNAase digestion of products from thin layers was carried out for 30 min at 0°C in 10 ~1 of a solution of 0.1 mg pancreatic RNAase/ml in 0.01 M-Tris.HCl, 0.001 M-EDTA, pH 7.5. Products were fractionated by ionophoresis on DEAE-paper using 7 O/oformic acid. Oligonueleotides were analysed by further digestion with pancreatic RNAase to completion (0.1 mg/ml in 0.01 M-Tris.HCl (pH 7*5), 0.001 M-EDTA in Lhe presence of 2 mg carrier tRNA/ml) followed by ionophoresis on DEAE-paper at pH 3.5. (d) Hybridization with 5 S RNA and preparation of antiparallel 5 S RNA 1 to 2 @Zi of purified 32P-labelled transcript (containing approximately 2 to 4 pg of carrier tRNA) was mixed with 4 pg of 5 S RNA (purified from oocytes or from tissue culture cells (somatic 5 S RNA) of X. Zaewis, Brown & Littna, 1966) in 100 ~1 of distilled water and heated to 100°C for O-5 min, rapidly cooled and then freeze-dried. The mixture was dissolved in 10 ~1 of 0.2 M-NaCl, 1 mivr-EDTA, 10 mM-Tris.HCl, pH 7.5, and incubated at 60°C for 15 min. 0.5 pg of T, RNAase in 10 ~1 of distilled water was then

706

G. G. BROWNLEE

ET

AL.

added to digest unhybridized RNA and the incubation continued for a further 30 min at 37°C. The ratio of T1 RNAase was thus one part of enzyme to about 10 to 20 parts of substrate. The reaction mixture was made up to 1 ml with 0.1 M-NaCl, 0.5% sodium dodecyl sulphate and T, RNAase was then extracted by shaking vigorously with phenol. The aqueous layer was applied to a Sephadex G50 column (see section (b) above). The material eluting in the void volume comprised about one-third of the original radioactive complementary RNA and included t.he 5 S RNA (32P-antiparallel)-5 S RNA hybrids. This material was concentrated by ethanol precipitation (see section (b) above) before analysis by complete digestion with T1 RNAase. Hybridization of the 3aP-labelled transcripts was also carried out on material purified without the use of carrier tRNA to show that it did not alter the specificity of the hybridization and the subsequent fingerprint that was obtained. (e) Polyacrylamide

gel electrophoresis

This was carried out in a multislot vertical slab gel apparatus as described by Brownlee (1972). A 2.5% acrylamide gel was prepared in 8 M-urea, 0.02 M-Trisacetate (pH 81), O*OOl M-EDTA (Ziff et al., 1973). The sample (a UTP-labelled transcript) was boiled for 1 min with 20 ~1 of the urea-containing buffer containing in addition 0.1% bromphenol blue and 30% sucrose before applying to the origin of a 20 cm x 20 cm gel. Ionophoresis was carried out at room temperature for 4 h at 250 V. 32P-labelled tRNA, 5 S RNA, 12 S mitochondrial and 18 S cytoplasmic ribosomal RNA, all from mouse myeloma cells (Brownlee et al., 1973) were run in parallel slots to serve as size markers. After radioautography, the transcript was divided into 2 size classes, 8 to 12 S and 12 to 18 S, and the RNA extracted by homogenization with 05 M-NaCl and precipitated by the addition of 2.5 vol. of 95% ethanol in the presence of 100 pg carrier tRNA. The RNA was purified from contaminating acrylamide by DEAE-cellulose column chromatography after digestion of the RNA with T, RNAase, as follows: 10 pg of T, RNAase (Sankyo) was added in 0.3 ml of 0.01 M-Tris.HCl (pH 7.5), 0.001 M-EDTA and after 30 min at 37°C the digest was loaded onto a 1 cm x 0.6 cm DEAE-cellulose column equilibrated in 0.1 M-Tris.HCl, pH 7.5. The column was washed with 1 ml of 0.05 M-NaCl before eluting with 1 m-NaCl, 7 M-urea, 0.01 M-Tris*IICl, pH 7.5. After precipitation of the oligonucleotides with 3 vol. of ethanol, the material was ready for fractionation by the two-dimensional system (see Plate II).

3. Results (a) Studies on the 5 S gene vegiole Is it oocyte or somatic type DlvAZ

Oocyte and somatic 5 S RNA differ in their antiparallel sequence at six positions as shown in Figure 1. Thus a T, RNAase digestion of such a sequence should give five distinctive oligonucleotides (Table 1) which differ according to whether the RNA is transcribed from somatic or oocyte type DNA. In addition, there should be 13 oligonucleotides common to both sequences. The design of the experiment was thus to establish which of the five pairs (Table 1) of distinctive oligonucleotides was present in a transcript

of 5 S DNA.

32P-labelled transcripts of 5 S DNA were prepared using both labelled CTP and UTP in separate experiments. The antiparallel RNA corresponding to the 5 S gene region was then purified from the spacer region by hybridization with oooyte 5 S RNA and digestion of unhybridized spacer transcripts with T, RNAase (see Materials and Methods). Plate I shows an example of a fingerprint of a complete T, RNAase digest in which transcription was carried out with labelled UTP. While much less complex then T, RNAase digests of total transcripts, the partly purified antiparallel complementary RNA fingerprint is still more complex than predicted for p”re

SEQUENCE Antiparallel 5’

707

OF 5 S DNA 5 S RNA

121 110 100 90 A-A-A-G-C-C-U-A-C-G-A-C-A-C-C-U-G-G-U-A-U-U-C~C-C-A.G.G-C-G-G-U-C-U-C.C-C~A~U80 70 60 50 C-C-A-G-G-U-A-C-U-A-A-C-C-A-C*-G-G-C~~-C-G-A-C-C~C-U-G~U-A~U-C-G-C-U-U.C-U~G-A. $4 4 j. CU G A 40

30

20

J C 10

G-A-U-C-A-G-A-C-G.A-G-A-U-C.A-G-G-C-A-C-U-U-U-C~A-G-G-G-~J-G-C-U~G-U.G-G~C~~.

1 G-U-A-G-G-C-

3’

FIQ. 1. Antiparallel (or complementary) sequence of oocyte 5 S RNA. The numbering is the same as is used for oocyte 6 S RNA (Wegnez et al., 1972). Those bases differing in somatic 5 S RNA (Brownlee et rrl.. 1972) are shown by arrows.

antiparallel 5 S RNA (see Table 1) there being at least 21 strong spots and many minor ones. Most of the major products were eluted and sequenced as shown in Tables 2 and 3. Unique sequences were determined by the use of pancreatic and Uz RNAase digestions for all except the four largest T, end products (g18, g20 and g21). In the last column of Table 2, the results are compared with the predicted sequences (Table 1). All of the sequences common to the antiparallel sequence of both somatic and oocyte 5 S RNA were found. Oocyte-specific oligonucleotides (g10, g12,g15, g19, g21) were identified and unique sequences were obtained for all except g21 (Tables 2,3 and 4). TABLE

Predicted

T, end-products

common

1

to, or speci$c for, oocyte and somatic type RNA?Specific1

Common

GFI, ‘WJI, (WI C-G[G] A-G[A]-2

A-U-C-A-G[G] (a) A-U-C-G[G]

Oocyte (0) or somatic (S) 0 S

moles

C-C-G[U]

C-C-C-G[A] A-C-G[A] U-G[G]-2 moles, U-G[U] U-A-G[G] A-U-C-A-G[A] C-C-U-A-C-G[A] A-C-A-C-CLU-G[G] U-A-C-U-A-A-C-C-A-G[G] C-A-C-U-U-U-C-A-G[G] U-A-U-U-C-C-C-A-G[G] . . . A-A-A-G[C]

A-C-C-C-U-G[U] A-C-C-C-U-G[C]

0 S

U-A-U-C-G[C] (b) C-U-U-G[G]

0 S

C-U-U-C-U-G[A] (c) C-U-U-C-C-G[A]

0 S

U-C-U-C-C-C-A-U-C-C-A.G[G] U-C-U-C-C-C-A-U-C-CLA-A-c[u]

0 S

t The sequences in this Table were predicted from Fig. 1, allowing cleavage at all G residues. $ The bracketed oligonucleotides are sequences present in either oocyte or somatic antiparallel 5 S RNA but not in both. Nucleotides (a) to (c) are identified in the t,ext and their possible locations on a fingerprint, are shown in Plate I.

spot

A-U,

n-u,

A,-G

g22

Absent

LJ c G, C

C

A-C-U-U-G[U] C-U-A-U-A-G[Pu] U(A-C-U, A-A-C-C).9.G[Pu] C:-U-U-C-XT-GCPu] Compatl%le with C-A-C-U-U-U-C-A-G and in lowor yi&.i TJ-A-U-U-C-C-C-A-G Compatible with U-C.U.C-C-C-A.U.C.C.A~G[G] A-A-A-A-GlC]

C-C-G[U] C0CG[Pu] A-C-G[Pu] U-G[U] A-U-G[N], A- U-G[U]-1 U-A-GIUl C-C-U-Gk] A-U-C-A-G[Pu] A-C-A-C-C-U-GLPu] A -C-C-C-U-G[U] C-C-U-A-C-G[Pu] C-C-C-C-U-G[U] U-A-U-U-C[C]

i A-G[U]

A-G[C]

WC1 Wul

Deduction

The following additional symbols are used in the Table: Pu = pu’irm residue, N = any nucleotide. Where a residue is once underlined, it was a visual inspection of the radioautograph to OCCUT’ in twice molar .yield. Where twice underlined, the yield was e&mated aa 3 moles. When the hour residue of a pancrcatio produot in specified, e.g. A-A-C[C] of g18, this indicates that alkaline hydrolysis was carried out giving the results the nearest-neighbonr is shown by [N], e.g. A-O [N] of g20, this indicates that alkaline hydrolysis did not give Cp. Thus N is either A, G 011U. prtnoreatic products wem identified by their position on the fmmtjionation qstjem used (i.e. DEAE-paper ionophoresis at pH 3.5). “Absent” no product, OP only low yields of product,, worn present in the T, RNAase fingerprint. t Present in low yield. $ Addit~iord data. needed for the derivation of these sequences is given in Tablo 8. 3 Composition 01‘ sequence was estimated from a knowledge of position cm the fingerprint (Banger et al., 1965). In colnplex oligonucleotides the number of U residues could be estimated, t.lw n.yrnbuls Ii = 1, 2 01‘ 3 are used.

u-3

g21

GU,)G

u-3

A,-G

A-C, U-UT, U, G A-U, C A-C: c, u A-U[U], A-C

A-C Absent A-A-C[C], A-C[N] u-ut, .u A-ClN], C!, U

(2, u

:c A’-U

or

C, Gt C A-C Absent Absent Absent G,r: A-U A-C A-C[C], C A-WW, C c! a-U, G

G Abmnt Absent a A-U, Gt A-G

C,-G C,-G (A, C)c: U-G A -U-G U-A-G W,, U)G CC, A,, U)G TJ = 1 G--3, A, U)G (Cs--3, A, UK (G, UK4 U = 2, (C,U,)G (G A, UdG CC, A, UzP (‘2, As, Us+> TJ = 2

&O

!316 331’ 6518 gl9

g3 84 & 236 @ 88 .!@ @OX gllS gw g13P d4 816

@ A-G

KNRascr digcmtian [32P1UTP input

A-G

Products of pancreat,io [““P]UTP

A-G

g c

Composition or sequellee a

G

gl

(PM e I)

1)

where

only

by estimated nearest-neighshown. Wham Otherwise the indicates that

Y-53

Yes, oocyte

NO NO Yt3S Yes, oocyte Y0S

YOY Y&3 No No YW3 Yes Yes Yea NO No NO Yes Yi3S Yes, oocyte Yes NO You, oocyte

Whether predicted (see Table

SEQUENCE

OF TABLE

709

6 S DNA 3

Analysis of spots gl0 to g13 by U, ribonudectse digestion spot no.

[3aP]CTP input

gl0

U-G-A C-C-U-G, C-A, A C-C-C-U-G, A C-C-U-A[C]

gll g12 gl3

[=P]UTP

input

Deduction A-U-C-A-G[Pu] A-C-A-C-C-U-G[Pu] A-C-C-C-U-G[v] C-C-U-A-C-G[Pu]

U, RNAase digestion was carried out as described in Materials and Methods, on Tr end products from a thin-layer fingerprint of the same material as was used to prepare Plate I. Fractionation was by ionophoresis on DEAE-paper at pH 1.9. The composition of the products was determined from position and the sequence could be deduced in all products in the CTP input digest from this information and a knowledge of the further digestion products with alkali to determine the nearest neighbour. The estimated composition of g10 to g13 is needed (see Table 2) for the correct deduction of the sequence. The mobilities of the Ua RNAase products relative to Ap are as follows : C-A, 0.93; U-C-A, 0.52; C-C-U-G, 0.20; C&-U-G, 0.16; C-C-U-A, 0.39.

The source of spot g10 cannot be determined since it can derive from both common and oocyte specific sequences. A number of oligonucleotides were identified which are not antiparallel to either kind of 5 S RNA and they are considered to be contaminating transcripts of the spacer region. The relative molar yields of the products in Table 2 were not measured because of the presence of contaminants which would clearly distort the quantitations of the smaller gene products. But, relative yields were measured on more purified antiparallel 5 S RNA (our unpublished observations) and the results show reasonable agreement with the predicted yields (Table 1). For example, the yields of g10, gll, g12 were, respectively (the nearest integral numbers being shown in brackets) 1.6 (2), 0.8 (l), O-9 (1). Spacer contaminants were present in low yields. Thus 87, g8, g9 and g-14were, respectively, 0.1, 0*1,02 and 0.05. Faint spots (a to c) on t,he fingerprint corresponded to the expected location of somatic sequences (see Table 1). Since these regions of the ohromatogram had too few counts to sequence, it could not be determined whether they were contaminants or low yield somatic antiparallel sequences. A convenient way of finding even a small amount of a somatic-type sequence was to measure the relative yields of A-C-C-CU-GlJl] characteristic of oocyte DNA and A-C-C-C-U-G[C] characteristic of somatic DNA (see Table 1). As the oligonucleotide is identical and only differs in its 3’ nearest neighbour, this product (g12) was eluted from a [32P]CTP input experiment in which hybridization was carried out with somatic 5 S RNA to maximize the chance of selecting somatic antiparallel transcripts. After pancreatic RNAase digestion of g12 the yields of A-C, C and G were measured. The proportion of G relative to A-C[C] in this digest is then taken as a measure of the proportion of A-C-C-C-U-G[C] contaminating A-C-C-CLU-G[U]. If only the latter sequence is present no G should be labelled as it is followed by a U. The results (Table 4) indicate that only a small amount of G is labelled. Since this small amount of G could have been derived from other oligonucleotides of composition (A, Ca, U)G[C] contaminating 812, the figure of 3% is probably an upper limit of the amount of A-C-C-C-U-G[C] present. We conclude that the level of somatic type 5 S DNA sequences is less than 3% of the amount of oocyte coding sequences. In this experiment, in which an attempt was made to enrich complementary RNA 48

710

0.

G. BROWNLEE

ET-AL.

TABLE

4

Comparison of yields of A-C-C-C- U-Q[Q and A-C-C-C- U-Q[U] Products of pancreatib RNAase digestion A-C[C] G C

Radioactivity (ctslmin)

G/A-C

2274 34 1121

0.03

A transcript of 6 S DNA was prepared using [3ZP]CTP as input label. The antiparallel gene region was purified after an RNA to RNA hybridization with tissue culture (sonaatic) 5 S RNA as described in Materials and Methods. g12 (Table 2) was eluted from a T1 RNAase fingerprint of this material and digested with pancreatic RNAase and fractionated by ionophoresis on DEAEpaper at pH 3.6. The yields of the products were estimated by counting the paper in a liquid scintillation counter. The figures for A-C[C] and C both include the counts in A-C cyclic phosphate and C cyclic phosphate which were both present in appreciable yield as separate bands. Note that A-C[C] has two moles of labelled PO+ for every labelled PO1 iu C.

antiparallel to somatic 5 S RNA by hybridizing with somatic 5 S RNA, a fingerprint was obtained which was qualitatively similar to the one shown in Plate I. Thus, hybridization with somatic 5 S RNA resulted in the enrichment of oocyte type sequences. One antiparallel sequence is of additional interest since it may provide information on the 3’ end of the gene region. Spot g22 (A,G[C]) corresponds to the region antiparallel to the 3’ terminus of 5 S RNA (Fig. 2). Denis & Wegnez (1973) have found oocyte 5 S RNA with 1,2,3 or 4 U residues at the 3’ end. These residues appeared to label differently from the rest of the molecule, causing them to conclude that they were added after transcription. Our finding suggests that this sequence may be part of the gene itself as proposed in Figure 2. DNA

(

E

(6’) (3’)

-A-G-G-C-T-T-T-T-C-T-C-C-G-A-A-A-A-G-

(3’) (5’) .

5 S RNA

(6’)

-+---22----t -A-G-G-C-U-U-U,,,

FICA 2. Proposed sequence of 6 S DNA corresponding

(3’)

to the 3’ end of 6 S RNA.

(b) Sequences in the spacer region, (i) Complete Tl ribonuckase digestion The spacer region is defined as that part of the 5 S DNA other than the gene region. Thus, a T, RNAase digest of a total transcript should contain both the gene sequences already characterized and additional oligonucleotides derived from the spacer. A considerable part, although not all, of the spacer is known to be high in A.T base pairs and would be expected to give high A and U-containing transcripts. Also, as G residues are relatively infrequent, large T, end products should be present in a complete T, RNAase digest. Plate II shows such a T, RNAase digest of a total transcript in which homochromatography was used in the second dimension to aid in the purification of the larger end-products of T, RNAase digestion. The numbered

PLATE 1. Two-dimensional fractionation of a complete T, RKAaso digest of [32P]UTP-labolled antiparallel 5 & RNA prepared by hybridization with 5 S RXA, digestion with T, RNAase, the resistant fragments purified on Sephadex and then T, digestion completed. Fractionation in the first dimension (direction 1) was ionophoresis on 85 cm strips of cellulose acetate at pH 3.5 followed by (direction 2) fractionation by ionophoresis on DEAE-paper using 7% formic acid. Spots are identified in Table 2. Spots 4, 5 and 22 show the approximate position of C,-G, A-C-G and A,-G, respectively, these being labelled only with a CTP input label. The dotted circles (labelled a to c) indicate the approximate position of the predicted specific somatic T, end products (a to c in Table 1). (C-U)G (adjacent to spot S), U-TT-G (adjacent to b) and IT-U-U-G (at extreme top left) m-ore the only major products that were not &ted. [facingp. 110

PLATE II. Two-dimensional fractionation of a complete T, RNAase digest of a total of 5 S DNA labelled with [3ZP]UTP. Fractionation in the first dimension (direction ionophoresis on a 55 cm strip of cellulose acetate at pH 3.5 followed by fractionation in dimension (direction 2) by homochromatography on a commercial DEAE-cellulose (Macherie & Nagel) using a 5% 10.min hydrolyzed homomixture. Spots are identified

transcript 1) was by the second thin-layer in Table 5.

PLATE 111. Two-dimensional fractionation of a pcwtial T, RNAase digest (see Materials and Methods) of a transcript of 5 8 DNA labelled with [32P]CTP. Fractionation was the same as in Plate II except that a stronger homomixture and a home-made DEAE-cellulose thin-layer was used in the second dimension (see Materials ant1 Methods). Spots are identified in Tables 5 and 7. A small ‘t’ refers to end-products of T, RSAaso digestion, whilst a capital ‘T’ refers to partial products.

SEQUENCE

OF

5 S DNA

711

products were sequenced (see Table 5) from data derived from four separate experiments in which each radioactive triphosphate was used in turn as the only input label. An unusual feature of these sequences besides their high A-content, is their similarity to one another (Table 6). The molar yields of these spacer products were estimated relative to gene spots which are assumed to occur once per repeat. The gene sequences used (Table 6) are numbered on Plate II as g 11 and g2 1. Table 6 shows that the yield of spacer oligonucleotides is not integral but varies over a wide range. Several products (t3, t7, t10, tll, t12) occur in more than one copy per gene, there being as many as five or six copies of tll, while others (tl, t4 t5, t6a and b) occur in less than one copy per gene. These results were obtained with four separate preparations of 5 S DNA and were independent of the label used in the transcription. We conclude that there are repeated sequences within a single spacer unit. The spots in less than molar yields are derived presumably from some spacers but not others, a fact which confirms previous demonstration of heterogeneity within 5 S DNA (Brown & Sugimoto, 1973a). The total number of residues accounted for by the products in Table 6, excluding t9 which seems unrelated in sequence to the other spots, may be deduced from a knowledge of the length and yield of the oligonucleotides. These sequences alone comprise about 230 residues per repeat which is about half the length of the high A + T part of the spacer (see Pig. 3). This estimate does not take into account differential losses of the larger relative to the smaller oligonucleotides (Brownlee t Sanger, 1967). As the yields of spacer products (Table 6) were estimated relative to, on average, smaller gene specific products, 230 residues must be a minimum value. Other A-rich sequences, clearly related to the above space sequences, occur in low yield. (This was not measured because they occur in positions on the fingerprint where they were not well resolved.) These include A,-U-G[A], A,-U-G-[A], (A&-C, U)G, (A,-C, U-C)G and A-A-C-A-G[C]. Many smaller oligonucleotides with U-rich, e.g. U-U-G[A] or A-rich, e.g. A-A-A-G occur in high yield but cannot be included in the quantitative analysis of the A + T spacer as they could not be definitely assigned to this region of the DNA. For all these reasons the 230 residue number is a minimum value and it is possible that the entire A + T region consists of sequences related to the ones shown in Table 6. (ii) Partial T, ribonuclease digestion The quantitative information presented in the previous section suggested the presence of reiterated sequences within a single spacer, These could be present as a continuous reduplicated sequence as occurs in many satellite DNAs (Southern, 1970; Fry et al., 1973) or the reiterated sequences might not be adjacent, We have used partial T, RNAase digestion in an attempt to isolate larger spacer sequences to test these possibilities. Plate III shows a fingerprint of such a digest. The characteristic T, end products are present in addition to some larger partial products moving more slowly in the second dimension of the fractionation system. Many spots were eluted and analysed (see Materials and Methods) but could not be sequenced uniquely because they were mixtures or because the data were incomplete. Nevertheless, results for three partial products, Tl-T3, were clear and are summarized in Table 7. Tl is a dimer of t7, while T2 is a trimer 45 residues long. T3 represents an overlap of t7 with t2, but as t2 is itself incompletely sequenced (see Table 5), the data were insufficient to establish the order of the overlap in T3. This partial product is of particular interest, however, in that it links a product present in high yield (t7) in the spacer

spot no.

1.2 o-9 1.0

1.0 1.0

1.0 1.0

A,-U[Cl A,-WI A,-CWI

&.-WI &-WI

&-WI &-WI

A&N1

A s-W7 A,-WI

&-WI A,-C[NI

t2

t3

t4

t5

t6

t7

U

2.0

1.0 1.0 0.9

1.0

O-6

1.0

2.0

Yield$

U

CTP

&CCW -%CENl

Input

t1

(Plate II)

u-u

A,-WJI &WI u

U u-u

As-UPJI A,-WJI 1 A,-WJI &WJl

U u-u

A,-WJI &-WJI

U u-u

&-W3 A,-WJI

A,-U[N] C U u-u

C U u-u

A,-UN A,-WJI

U u-u

(0.5)

Trace 1.0 1.0 1.3

2.3

(1) (1)

1.0

Trace

2.3

Trace 1.1 1.0

I.7

Trace 1.0 1.0

o-7 0.7

1.0

Tr&Ce

1.2 0.9 0.8 24

TraC0

0.7

2.0 0.7

of pancreatic UTP Yield

AI-WJI A,-C[Nl

Products Input

TABLE

5

U u-u u-u-u

U u-u u-u-u

U u-u u-u-u

U u-u u-u-u

U u-u U-G

U u-u

C

(1) Trace Trace

Trace

Trme

(1)

(1) Trace Trace

(1) Trace Trace

(1) Trace Traoe

(1) Trace

RNAase digestion Input GTP Yield A,-C As-C G c U A,-U A4-C A,-C G C U A,-U A,-C G C As-U As-C G C A,-U As-C G U A,-U A,-U A,-C G c U A,-C As-C G C

Input

ATP

(0.5) (1) (1) (O-5) (1) (1) (1) (1) (1)

(2)

(2-3) (O-5) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1)

(2)

(1) (1)

Yield . . . , A,-C-U

5

A,-C-U-U-U)A,-C-C-U-U-G[A]‘b,

. . . , A,-C-U

Deduotion

A,-U-U-A,-C-U-U-U-G[A]‘&’ A,-U-C-A,-U-U-U-U-G[A]‘a’

AG-U-U-U-A,-C-U-U-U-G[A]@)

(A,-U-C,

(A,-C-U

Analysis of T, ribonuclease end products of spacer regiont

. . .) C-G[A]l

A-U[C] A-C[N] U u-u

&-WI &-WI

&-WI

U

t9

t10 t11

t12

tr3

1.0 2.0 1.0 Trace

1.0 1.4 Trace

U

1

&-QVJI u-u

U

:*-C[U] u-u &J, 1

c U U-U A-U[N] u-u A-C[U] U

&-WJI

3.3

1.0

1-o 2.0

2.0

1.0

0.8 Trace 1-o Trace 0.8 0.8 Trace Trace (1) Trace Trace

Aa-WNI C

&QFl

G

&-WI

G

A.dWI &-WI

(1)

&-WI U u-u u-u-u U u-u u-u-u

A-G[A] A-C[A] C U

AI-C C

A-G[N]

A-G[G]

(1)

(1) (1)

(1) (1)

(1) (1)

(2) (1)

(1)

(1)

(1) A-C-U, U-U-C,

U-U-U-U-C-As-G[U]Ij

A&-U-U-U-G[A]

A,-C-G[U] A,-C-U-U-U-G[A]

(A-U-C,

A,-C-U-U-C-U-C-A-G[G]‘C’

A-C-A)A-G[A]B

this oligonucleotide

on DEAE-paper

using 7% formic

acid showed U = 4, thus resolving

..-

_ -

.-

-

in the partial

--- ---

sequence given. tl

of three or four U residues. included

the ambiguity

7 In these oligonucleotides, there was insufficient data to deduce a sequence, Not all residues are necessarily may be a mixture, as there is a low yield of U in the input ATP experiment.

i Rerunning

$ The sequences shown for t3-t8 cannot be deduced uniquely from the information in this Table, with the exception of t4. The following additional methods were used to resolve the sequences: (a) Partial digestion with Us RNAase (see Materials and Methods); t7 gave C-U-C-A,[C], t6 gave both C-U-U-U-G and U-U-U-U-G, the former in higher yield and t3 gave (C,, U,)G. (b) Seoondary splitting (Brownlee & Sanger, 1967) with T1 RNAase gave A,-C-U-C and As-C-U-U-U-G with t7 and (As-U-C, Ac-C-Us) and As-C-C-U-U-G for t2. (c) Partial pancreatic RNAase digestion (see Materials and Methods) gave A,-C-U-U-C, thus resolving the sequence.

$ Yields were measured in a scintillation counter and are corrected for the number of phosphate residues labelled. Where the figure is enclosed in brackets, e.g. (0.5), or “Trace” is indicated, the estimate was made by visual inspection of the radioautograph. U-U and U-U-U are partial pancreatic digestion products oonunouly prosent in trace amounts.

7 Panoreatic digestion products were identified by position on ionophoresis on DEAE-paper at pH 3.6. Larger products, e.g. As-C, As-U, As-U, were tentatively identified on this system but were confirmed by rerunning on DEAE-paper using 7% formic acid. Most of the products were eluted and treated with alkali to determine the nearest neighbour which is shown in square brackets [I. Where [N] is shown, this indicates that the 3’ terminal nucleotide was unlabelled in the alkaline hydrolysate. Thus for the iqmt label spe,eciJed there is no transfer of the label and [N] is any base other than that used for the input label, e.g. in tl, A,-C[N] in the column headed “Input CTP”, [N] is either U, G or A but not C.

&-WI

U u-u

ts

714

G. G. BROWNLEE TABLE Comparison

of sequences

(A,-C-U

. . . , A,-C-U

AL.

6 yields

of some spacer products Average yield (range)

Sequence

spot t1

and molar

ET

. . .)C-GrAJ

0.1 (0.1-0.2)

t2

. , . , A,-C-U

C-C U-U-G[A]

0.4 (0.2-0.6)

t3

C-C U-U-G[A]

i.4 (1.1-1.7)

t4

C-U

U-U.G[A]

0.6 (04-0.7)

t’5

C-U

U-U-G[A]

0.6 (04-0.9

t6a tGb

C-U U-U-G[A] U-U U-CJ-G[A]

t7

C-U

t10

U-U-G[A]

3

1.6 (1.2-l%) 3.4 (2.4-4.0) 5.3 (4.8-5.9)

i

t11

C-U

t12

C-U L-J U-U-G[A] t-G[A]

U-U-G[A]

5.5 (3.7-7.0) 2.5 0.8 (Od-1.0)

t9

(A-U-C, A-C-U, U-U-C,

t13

U-U-U-U-C-A-A-A-G[U]

1.8 (1.2-2.1)

gll

A-C-A-C-C-U-G[G]

0.9 (0.8-1.0)

g21

U-C-TJ-C-C-C-A--U-C-C-A-G[G]

1.2 (1.0-1.4)

A-C-A

No. of measurements

8

with one present in low yield (t2). Thus T3 is likely to be derived from only half of the total spacer regions of 5 S DNA. We conclude that the A + T rich region of the spacer is constructed of closely related sequences and that these sequences can be repeated tandemly in a manner reminiscent of satellite DNA. (c) General remarks Conclusions on the spacer transcripts depend on several features of the transcript and the DNA itself. During the course of these studies it became apparent that 5 S DNA was not always pure. One important observation was that a restriction enzyme from Haemwhilus injluenzae (Hind III, Smith & NTathans, 1973) cleaves about 80 to 90% of 5 S DNA once each repeat at the junction of the A + T and G + C region (Carroll & Brown, unpublished observations). Most of the larger uncut molecules appear to be contaminants; this raised the possibility that the contaminant might have been the source of the presumed spacer sequences. This possibility was ruled out by isolating the repeat length fragments free of the contaminant by gel electrophoresis and demonstrating the spacer sequences in transcripts from this DNA. There is a possibility in the experiments so far described that the T, RNAase fingerprint (Plate II) might represent transcripts of selected regions of the H strand of 5 S DNA rather than being representative of the entire H strand. If this were so, different sized transcripts might give different fingerprints. In an attempt to exclude this possibility, experiments were carried out to fractionate transcripts from native 5 S DNA into size groups. The [32P]RNA transcripts from native 5 S DNA were fractionated by acrylamide gel electrophoresis in 8 M-urea. The gel was calibrated by using marker tRNA, 5 S ribosomal RNA, 12 S mitochondrial and 18 S cytoplasmic

A&U-C-A,-C-U,-G (t7) A,-C-U-C-As-C-Us-C-Us-G (t7) (+ some contaminants) A4-C-U-C-As-C-Ua-G (t7) (As-U-C, A4-CLUs)As-C-C-U-U-G (t2)

T, end products

As-U, G-A,-C,

A,-C, Aa-C[C]

G-Ad-C, A,-C, A,-C G-A,&?, A,-C, A,-C C

digestion

Sequence deduced A4-C-U-CLAs-C-Us-G-A,-C-U-C-As-C-U-U-U-G A,-C-U-C-A,-C-UB-G-A,-C-U-C-As-C-U3-G-A,-C-U-C-U-AsC-U,-G [A,-C-U-C-A,-C-Us-G][(As-U-C, A,-C-U,)A,-C-C-U-U-G]

products of T, ribonuclease

Pancreatic RNAase end products (CTP input)?

of partial

7

The data in this Table were obtained independently from both input CTP and input UTP experiments. The T, end products were fractionated on homochromatography and characterized by further pancreatio RNAase digestion (see Materials and Methods and Table 2) whilst the pancreatio RNAase end products were separated by ionophoresis on DEAE-paper using 7% formic aoid and analysed further by T, RNAase digestion (see Materials and Methods). t Mononucleotides are not included.

T3

Tl T2

spot

Analysis

TABLE

716

G. G. BROWNLEE

ET

AL.

ribosomal RNA (see Materials and Methods). The complementary RNA was polydisperse in size and a complex band pattern was present from 4 S to 18 S in size. The average size was variable in different experiments. In one experiment with a [32P]CTP input label, it was about 600 residues, while in another experiment with UTP it was about 1000 residues. In this latter experiment with UTP, T, RNAase fingerprints of the larger 12 to 18 S RNA fraction isolated from gels was identical to those of the smaller, 8 to 12 S RNA fraction as well as to the total unfractionated material. The 12 S to 18 S RNA fraction includes a range of sizes from 1000 to 2000 bases, which is on average twice as long as the repeat monomeric unit of 5 S DNA. We find no evidence for selective transcription of the 5 S DNA although these experiments cannot rule out the possibility conclusively.

4. Discussion (a) 5 i3 gene region Sequence studies of 5 S DNA have clarified some of the details of its structure. The gene region seems to code only for the 5 S RNA synthesized by oocytes. If less than 3% of the transcripts were somatic in type, we would not have detected them. The estimated number of total 5 S genes is 24,000 (Brown et aZ., 1971) and we can now say that most of these must code for oocyte 5 S RNA. One oocyte synthesizes and accumulates more than 200 times as many ribosomes as the most active somatic cell in a given period of time. This synthesis is made possible by massive amplification of genes for 18 S and 28 S rRNA (Brown & Dawid, 1968; Gall, 1968). Oocyte-type 5 S genes however are not amplified but are present in thousands of copies in the germ line as well as in all somatic cells. Evolution has provided two different solutions for the massive synthesis of ribosomes in oocytes. Amplification of 18 and 28 S rRNA genes in oocytes provides extra templates for them, and these genes are discarded at meiosis (Brown & Blackler, 1972). No such mechanism occurs for 5 S genes used during oogenesis. They are present but presumably not used in all somatic cells and therefore a permanent part of the animal’s genome. Another set of 5 S genes, under separate control, are expressed in somatic cells. It is likely, though not demonstrated conclusively here, that the 5 S DNA used for these experiments is totally devoid of the somatic sequences. A by-product of the sequence work was the confirmation of the structure of oocytespecific 5 S RNA by the determination of the sequence of the complementary T, RNAase end products. The results of this type of analysis fully confirm the structure of the oocyte 5 S RNA (Wegnez et al., 1972; Ford & Southern, 1973). They also provide clear evidence that E. coli RNA polymerase faithfully copies Xenopus 5 S DNA. (b) The 5 8 DNA repeat One 5 S DNA repeat consists of a single copy of the 5 S RNA gene which is about 121 base pairs and a spacer about six times as long with an overall average length of about 750 base pairs (Fig. 3). Its thermal denaturation profile and denaturation map (Brown et al., 1971) demonstrated that 5 8 DNA consists of two widely different regions of which about 60% denatures about 10 deg. C before the remaining 40% of the DNA. Denaturation mapping shows the simple alternating arrangement of these two regions (designated A + T and G + C in Fig. 3). A single repeat consists of one A + T region adjacent to one G + C region and this arrangement repeats

SEQUENCE -440 I

OF

717

6 S DNA f---270-

I

FIG. 3. Diagram of one average 5 S DNA repeat showing the internal repeats within the A + Trich region. The sequence of one of the most abundant sequencas (t7, Plate II, Table 6) is given. The dotted line dividing the oligomer is meant to suggest that this sequence itself was derived from a duplication. The 6 S gene comprises one-half to one-third of the G + T-rich region and is placed arbitrarily as its exact location is unknown.

without interruption of any other genes for as many as 1000 times on any one chromosome (Pardue et al., 1973)), The gene region comprises about one-half to one-third of the base pairs within each G + C region. The A + T rich part of the spacer has been found to consist of extensive repeating sequences. Twelve different but closely related oligonucleotides, on average 15 residues long have been identified by T, RNAase sequence analysis (Table 6). By comparing the molar yield of these spacer sequences with oligonucleotides known to be derived from antiparallel gene sequences and therefore presumed to be present in a single copy per repeat, the average number of each of these has been determined (Table 6). They range from less than one per repeat (tl and t2) to several copies per repeat (t7, t10 and tll). These differences reflect the sequence heterogeneity known to occur within the spacers of 5 S DNA (Brown & Sugimoto, 1973a). The summation of these identified sequences is enough to comprise about 230 residues per repeat which is half of the A + T rich region. This is a minimum estimate. Smaller oligonucleotides having more G substitution than those given. in Table 6 were found and partly sequenced. Alth ough undoubtedly related to those given in Table 6 they could not be quantitated accurately and were therefore omitted from the Table. These added to the ones in Table 6 may well account for the entire A + T-rich part of the spacer. Additional evidence that at least some of these oligomers are adjacent comes from partial T, digests (Plate III, Table 7) in which dimers and trimers of one 15-mer were identified. We conclude that these repeating oligomers are short internal repeats within a larger repeat. Indeed these 15mers themselves may be duplications of an AA c with something like the structure A,CC T T. TT Whatever the function of spacer DNA sequences, it is not likely to depend on the exact sequence or length of the spacer region. This has been concluded by comparing the spacers in the 5 S DNA of X. laevis with those of X. mulleri (Brown & Sugimoto, 1973a). The spacer sequences of these two kinds of 5 S DNA are different enough so that they do not hybridize, and X. mu2leri 5 S DNA spacers are on average more than ancestral oligonucleotide

718

G. G. BROWSLEE

ET

AL.

tnice as long as t.hose of X. luevis. Possibly 5 S DNA spacer is required as a means of maintaining the relative homogeneity of the sequence coding for 5 S RNA while still allowing many gene copies. It may do this by allowing frequent crossing-over at meiosis (see Brown & Sugimoto, 1973u,b). Clearly this recombination would be enhanced by the kind of short reiterated sequences reported in this paper which are characteristic of the high A + T spacer of 5 S DNA. REFERENCES Broxvn, D. D. & Blackler, A. W. (1972). J. MOE. Biol. 63, 75-83. Brown, D. D. & Dewid, I. B. (1968). Science, 160, 272-280. Brolvn, D. D. & Littna, E. (1966). J. Mol. Biol. 20, 95-112. Brown, D. D. & Sugimoto, K. (1973a). J. Mol. BioZ. 78, 397-415. Harbor Symp. Quant. BtiZ. 38,501-505. Broxvn, D. D. & Sugimoto, K. (1973b). ColdSpring Brown, D. D. & Weber, C. S. (1968). J. Mol. BioZ. 34, 661-680. Bronm, D. D., Wensink, P. C. 8i Jordan, E. (1971). Proc. Nat. Acud. Sci., U.S.A. 68, 3175-3179. Brovmlee, G. G. (1972). Determination of Sequences in RNA, Part I of vol. 3 of Laboratory Techniques in Biochemistry and Molecular Biology, (Work, T. S. & Work, E., eds), North-Holland, Amsterdam. Brownlee, G. G. & Sanger, F. (1967). J. Mol. BioZ. 23, 337-353. Brownlee, G. G. & Sanger, F. (1969). Eur. J. B&hem. 11, 395-399. Bro\vnleo, G. G., Cartwright, E., McShane, T. & Williamson, R. (1972). FEBS Letters, 25, 8-12. Brovmlee, G. G., CartJvright, E. M., Cowan, N. J., Jarvis, J. M. & Milstein, C. (1973). Nature h’ew BioZ. 244, 236-240. Burgess, R. R. (1969). J. BioZ. Chena. 244, 6160-6167. Denis, H. & Wegnez, M. (1973). Biochimie. 55, 1137-1151. Ford, P. J. & Southern, E. M. (1972). Nature New BioZ. 241, 7-12.

Forget, B. G. & Weissman, S. M. (1967). Science, 158, 1695-1699. Fry, K., Poon, R., Whitcome, P., Idriss, J., Selser, W., Mazrimas, J. & Hatch, F. (1973). Proc. hTat. Acud. Sci., U.S.A. 70, 2642-2646. J. G. (1968). Proc. brat. Acud. Sci., U.S.A. 60, 553-560. l?ardue, M. L., Brown, D. D. & Birnstiel, M. L. (1973). Chromosoma, 42, 191-203. Reeder, R. H. (1973). J. Mol. BioZ. 80, 229-241. Sanger, F., Brovmlee, G. G. & Barrell, B. G. (1966). J. Mol. BioZ. 13, 373-398. Smith, H. 0. & Nathans, D. (1973). J. MoZ. BioZ. 81, 419-423.

Gdl,

Southern, E. M. (1970). Nature (tin&n), 227, 794-798. Wegnez, M., Monier, R. & Denis, H. (1972). FEBS Letters, 25, 13-20. Ziff, E. B., Sedat, J. W. t Galibert, F. (1973). Nature New BioZ. 241, 34-37.