Cell, Vol. 12, 183-189,

September

1977, Copyright

0 1977 by MIT

Discrete Sizes of Replication Drosophila Cells Alan B. Blumenthal and Elizabeth Laboratory of Radiobiology University of California San Francisco, California 94143

J. Clark

Summary The size of DNA replication intermediates from Drosophila cells pulse-labeled with 3H-thymidine for 30-120 set was determined by electrophoresis in formamide-polyacrylamide gels. Replication intermediates were formed in three discrete size classes, with median lengths of 61, 125 and 240 nucleotides. Replication intermediates in the 125 nucleotide size class occurred most frequently. Two of the three size classes may contain discrete species of replication intermediates about 90-400 nucleotides long. The data also suggested that some larger replication intermediates accumulate in pulse-labeled cells. We concluded that 61 nucleotide molecules give rise to 125 and 240 nucleotide molecules, which then form high molecular weight DNA. Mechanisms for forming these replication intermediates are discussed. Introduction Newly synthesized DNA from Drosophila cells in culture appeared as molecules of about 120 or 225 nucleotides (3.8 or 4.8s) when analyzed on alkaline velocity gradients (Blumenthal and Clark, 1977). We demonstrated that the 120 nucleotide replication intermediates were precursors of the 225 nucleotide replication intermediates, which then gave rise to high molecular weight nuclear DNA. Those studies showed broad distributions of replication intermediates around the average sizes of 120 or 225 nucleotides. The resolution of velocity sedimentation analysis was not sufficient to determine whether the broad distributions observed on gradients were continuous or whether they were composed of discrete molecular species. We performed high resolution electrophoretic analyses of replication intermediates to answer this question and to define the mechanism of replication intermediate synthesis. Results Isolation of Replication Intermediates Replication intermediates were isolated from Drosophila cells in culture that were pulse-labeled with 3H-thymidine for 30, 60 or 120 sec. The 3H-labeled DNA molecules were separated by size on alkaline sucrose velocity gradients. Figure 1 shows that the peak of replication intermediates (4-5s) migrated

Intermediates

in

about 40% of the gradient length. The gradient fractions were pooled into three regions (A, B and C), which contained molecules from 0 to 2-35 (A), from 3 to 6s (B) and from 6 to greater than 10s (C). The A region contained 14% of the 3H counts found in the B region of this gradient (Figure 1); the average value in several gradients was 18 +- 3% (n = 7). DNA from the pooled regions of replicate gradients was combined and ethanol-precipitated. The A region material contained most of the remaining undigested cell proteins (90% of the protein to acid-soluble molecules. The 3HDNA was sedimented through an alkaline (0.2 N NaOH, 0.6 N NaCI) 5-20% sucrose gradient, and fractions were assayed for acid-precipitable 3H-DNA. Fractions from regions A, B or C were pooled together and combined with those of replicate gradients for further analysis.

61 + 1.5 nucleotides (n = 10). This is shown in the analyses of molecules labeled for 30, 60 and 120 set (Figures 3A, 3B and 3C). The band width of this distribution is about 2-3 times larger than that of a homogeneous marker fragment of comparable size and suggests the presence of molecules 50-70 nucleotides long. The A region also showed a bimodal distribution of molecules 110 and 140 nucleotides long, a minor component at about 250 nucleotides and a large amount of DNA at approximately 1500 nucleotides. The occurrence of 1 lo250 nucleotide molecules in the A region was probably due to contamination by B region molecules during collecting and pooling of fractions from several gradients for each electrophoresis sample. The component in these gels with a low mobility (about 1500 nucleotides) may have been due to aggregation of replication intermediates with proteinase K molecules and remaining cellular proteins. We have observed (unpublished observations) that replication intermediates and marker DNA fragments had a low mobility in these gels in the presence of cell lysate proteins. The absence of a low mobility component in gels of B region material, which contains little if any contaminating protein, also suggests that this component in A region DNA is due to aggregation of proteins and DNA. The 6 region replication intermediates, shown in Figure 4, contained 61 nucleotide molecules, a

1

Figure 2. The Electrophoresis

I

I

I

I

I

10

8

6

4

2

MOBILITY

(cm)

Mobilities of Viral in Polyacrylamide

DNA Fragments Separated Formamide Gels

0

by

3H-labeled 4Xl74RF DNA and SV40 DNA were digested with Hae Ill or Hind Ill, denatured, and electrophoresed in 5% polyacrylamide and 96% formamide gels. The graph shows the mobilities of these fragments against the log of their lengths. Length assignments were taken from the work of Carroll and Brown (1976). (A) Hae Ill digest of SV40 DNA; (0) Hae Ill digest of 4X174 DNA; (0) Hind Ill digest of SV40 DNA. The SV40 and 4X174 fragments in this example were run on different slab gels. Prints of the fluorograms (below), showing the electrophoretic patterns of DNA fragments, were overexposed to accentuate the smaller fragments. The arrow indicates the 265 and 290 nucleotide 4X174 fragments, which are well resolved in shorter exposures. The top photograph (A) is slightly shorter than the scale given on the graph.

broad distribution of 90-150 nucleotide molecules, and a shoulder or secondary distribution from about 170-300 nucleotides. The median values of the latter distributions were 125 + 1.7 (n = 21) and 240 -c 8.5 (n = 10) nucleotides. Analyses of C region replication intermediates showed one major component which migrated at about 1700 nucleotides (results not shown). When a large amount of C region DNA was run longer (1.7 times) on a 40 cm slab gel to resolve molecules of low mobility, several components of 300-1900 nucleotides were resolved, as shown in Figure 5. A

Drosophila 185

Replication

Intermediates

I:

larger component observed, which origin.

A

% c

0 0”s

4

ii

‘i

8

i t

S$ ii

L 8

4

6

0

2

CENTIMETERS FROM ORIGIN

(>2500 nucleotides) was also migrated about 1 cm from the gel

Heterogeneity of Replication Intermediates Bands of molecules having widths similar to those of marker DNA fragments were observed in many of the fluorograms within the broader size classes of replication intermediates. For example, there is a discrete band in Figures 3B’ and 3C’ at 110-112 nucleotides and another at about 147 nucleotides. Several faint but discrete bands were also visible in the fluorograms shown in Figures 4A’-4C’. Their positions appeared to correspond to minor peaks and shoulders seen in the microdensitometer traces (Figures 4A-4C). The positions of these bands were measured directly on the fluorograms because of the loss of resolution in photographic prints or microdensitometer traces. Molecules from 85-420 nucleotides long were observed several times, with 113 nucleotide molecules appearing most frequently. These measurements suggested the following sizes, in nucleotides, of replication intermediates [range of observed sizes (number of observations)]: 93 [85-99 (4)], 113 [l lo119 (lo)], 130 [130-132 (3)], 152 [147-159 (6)], 174 [168-178 (3)], 252 [233-280 (4)], 329 [315-352 (4)], and 395 [370-420 (3)]. The heterogeneity of B region replication intermediates was also observed by electrophoresis in 5 and 8% polyacrylamide tube gels. This method provided direct quantitation of the 3H counts in replication intermediates. The results, shown in Figures 6A and 6B (8% gels) and 6C (5% gel), consistently demonstrated a minimum of three well separated components in the lower molecular weight region (for example, fractions 20-40 in Figure 6B) and possibly several higher molecular weight components (for example, fractions l-10 in Figure 6A; fractions l-20 in Figure 6B). Discussion

B’ C’ Figure cation

3. Electrophoretic Intermediates

Analyses

of A Region

Drosophila

Repli-

3H-labeled DNA replication intermediates from the A region of the sucrose velocity gradients (see Figure 1) were analyzed by electrophoresis in 5% polyacrylamide and 98% formamide gels. The densitometer tracings of gel fluorograms are shown above: (A) 30 set pulse label; (B) 60 set pulse label; and (C) 120 set pulse label. The numbers above the tracings indicate molecular lengths in nucleotides. The photographs (below) show the fluorograms of the electrophoretic gels scanned in A, B and C. A’ corresponds to A, and so on.

Sizes of Replication Intermediates Drosophila replication intermediates, analyzed by polyacrylamide gel electrophoresis, were distributed in three discrete size classes-61, 125 and 240 nucleotides. We previously derived similar size estimates (120 and 225 nucleotides) for the larger two size classes from velocity sedimentation in alkaline sucrose gradients (Blumenthal and Clark, 1977). Those results also suggested the existence of 50 nucleotide replication intermediates. The distributions of 3H counts in velocity gradients and electrophoretic gels (corrected for the relative lengths of these molecules) gave the approximate proportion of 2:5:1 for the numbers of replication intermedi-

Cell 166

ates in the 61, 125 and 240 nucleotide size classes. Analysis of replication intermediates from human lymphocytes (Tseng and Goulian, 1975), mouse cells (Gautschi and Clarkson, 1975) and Chinese hamster cells (Huberman and Horowitz, 1974) by

gel electrophoresis or velocity sedimentation also shows loo-120 nucleotide intermediates, which correspond to the 125 nucleotide molecules from Drosophila cells. The electrophoretic analyses also showed that certain sizes of replication intermediates were less abundant. We observed few molecules shorter than 40 nucleotides (Figures 3 and 4), although the gels were capable of resolving marker molecules of about 30 nucleotides. If replication intermediates of this size exist, they were lost during isolation or occurred infrequently. There was also a lack of replication intermediates of 70-100 nucleotides (Figure 4). The 125 and 240 nucleotide classes may contain discrete molecular sizes ranging from about 90 to about 400 nucleotides. The most frequently observed discrete band of replication intermediates was about 113 nucleotides long.

Sequence of Synthesis of Replication Intermediates

J, 8

6

4

2

0

Our finding that replication intermediates occurred in three size classes indicates that their synthesis from l-250 nucleotides is discontinuous. A similar conclusion was derived from studies of the size distributions of replication intermediates in human cells (Tseng and Goulian, 1975; Kuebbing, Diaz and Werner, 1976). The results presented in our previous studies of pulse-labeled Drosophila cells (Blumenthal and Clark, 1977) indicate that replication intermediates are produced in sequential steps and that the various sizes accumulate during production. Specifically, the average size of replication intermediates increased from about 3S (66 nucleotides) to about 5S (250 nucleotides) during 120 set of synthesis. Thus the 61 nucleotide replication intermediates seen on the gels in our, present study may be precursors of the 240 nucleotide intermediates. Sedimentation analysis of data from our earlier pulse and pulse-chase experiments indicated that the 125 nucleotide molecules are precursors of the 240 nucleotide molecules, and that both classes of molecules comprise a steady state precursor pool for the synthesis of high molecular

CENTIMETERS FROM ORIGIN Figure 4. Electrophoretic mediates

Analyses

of B Region

Replication

Inter-

Replication intermediates from the B region of sucrose velocity gradients were analyzed as in Figure 3. The densitometer tracings and corresponding photographs show replication intermediates from 60 set pulses (A and B) and a 120 set pulse (C). (B) and (C) samples were obtained from the same experiment, and (A) was obtained from a separate experiment. The photographs (A’) show two different exposure times for this sample. The lower fluorogram of the pair shows bands at 60 and 370 nucleotides. The samples shown in fluorograms B’ and C’ were concentrated toward the edges of electropherograms. This nonuniform distribution across the gel was exaggerated by the contrast enhancement used in photographic printing.

Drosophila 107

Replication

Intermediates

1

I

12

10

8

I

I

I

6

I

4

I

I

I

2

CENTIMETERS FROM ORIGIN

Figure 5. Electrophoretic mediates

Analysis

of C Region

Replication

Inter-

See legend to Figure 3 and text for additional information. Several cracks in the gel caused downward deflections in the microdensitometer tracing (that is, at 420, 770 and between 1500 and 1700). The constriction in the fluorogram was due to the large amount of higher molecular weight material on the gel.

weight chromosomal DNA. Studies in other systems have also demonstrated that the 120 nucleotide intermediates are precursors of higher molecular weight DNA (Huberman and Horowitz, 1974; Tseng and Goulian, 1975). The occurrence of intermediates larger than 300 nucleotides in Drosophila (Figure 5) and in mammalian cells (Huberman and Horowitz, 1974; Tseng and Goulian, 1975; Kowalski and Cheevers, 1976) suggests that there are additional steps in the synthesis of high molecular weight DNA from smaller intermediates. Thus the 61, 125 and 240 nucleotide molecules that we found in Drosophila cells may be sequential intermediates in the synthesis of high molecular weight DNA. Formation of Replication Intermediates There are at least four possible mechanisms for the formation of DNA replication intermediates. The first two mechanisms involve enzymatic activities not directly related to discontinuous DNA replication. One of these mechanisms is that endonuclease cleavage of newly synthesized DNA during the isolation procedure creates replication intermediates. In our earlier studies, however, we demonstrated that a variety of stringent cell lysis conditions produced the same size distributions of repli-

I

I

1

20 FRACTIONS Figure 6. Electrophoretic mediates in Tube Gels

/

40

Analysis

FROM

60

00

ORIGIN

of B Region

Replication

Inter-

B region replication intermediates were electrophoresed through 8 or 5% polyacrylamide and 98% formamide tube gels. The positions of DNA molecules were determined by the 3H counts in fractions cut from the gels. (A) and (8) show the results from 8% gels run for different times, and (C)shows the result of the sample run on a 5% gel. The middle of the bromophenol blue markers migrated 75, 112 and 127 fractions from the origin in (A), (5) and (C). respectively. Recovery of 3H counts from these gels was about 40%.

cation intermediates on velocity gradients, and that endonuclease activity (assayed by nicking of form I SV40 DNA) was negligible under these conditions (Blumenthal and Clark, 1977). We cannot exclude the possibility that specific endonucleases, which are tightly bound to replicating DNA [for example, the cis A protein of $X174 (Scott et al ., 1977) or the site-specific nuclease of SV40 (Kasamatsu and Wu, 1976)], cleave these molecules. This mechanism

Cell 188

could not, however, easily explain the “chase” of replication intermediates into high molecular weight DNA. Replication intermediates could also be formed by the process of DNA repair. It has recently been proposed that incorporation of uracil bases into newly replicated DNA, in place of thymine bases, caused the accumulation of 4-5s replication intermediates (Okazaki fragments) in bacterial cells (Tye et al., 1977). The repair of the incorporated uracil bases could result in endonucleaseor alkali-catalyzed cleavage of phosphodiester bonds and in the transient production of fragments from newly synthesized DNA. A similar process could occur in eucaryotic cells (Lindahl, 1975), explaining the production of some replication intermediates. Because uracil would be randomly substituted in the place of thymine, however, we would not expect these fragments to occur in discrete size classes. In the second two mechanisms of formation of replication intermediates, intermediates are part of normal discontinuous synthesis of DNA, as first proposed by Okazaki et al. (1968). The several sizes of replication intermediates may occur because the DNA polymerase either stops or pauses at intervals of about 60, 125 and 240 nucleotides. These stops or pauses could be directed by the interaction of the DNA polymerases with nucleosomes. This model predicts that the sizes of replication intermediates should correspond to some structural parameters of nucleosomes. A Drosophila nucleosome contains about 200 nucleotides of DNA after limited micrococcal nuclease digestion (A. Worcel, personal communication) and is composed of a 60 nucleotide spacer region and a 140 nucleotide core particle (for example, Compton, Bellard and Chambon, 1976). The 61 nucleotide intermediates correspond with the spacer region, but none of the other replication intermediate lengths can be conveniently correlated with parameters of nucleosome structure. However, the lengths of DNA associated with nucleosomes may vary (Garrard and Todd, 1977). It is also possible that nucleosome structure is altered in the region of DNA replication forks (for example, as half-nucleosomes; Weintraub, Worcel and Alberts, 1976), resulting in nucleosomes containing DNA lengths different from those found in nuclease digests of nonreplicating chromatin. A fourth possible explanation for the formation of replication intermediates is that DNA polymerizing enzymes measure out the lengths of these molecules. The DNA polymerase(s) in vivo may be selflimited to synthesizing 60 nucleotide molecules. These could then be joined by DNA ligase to form larger replication intermediates and high molecular weight DNA. The necessity for synthesizing one daughter DNA

strand in the 3’ to 5’ polarity and the inability of DNA polymerases to carry out this reaction argue for the direct role of replication intermediates in discontinuous DNA synthesis, as in our last two suggestions above. The three size classes of replication intermediates which we have observed and the suggestion of discrete sizes of molecules within these classes imply a mechanism specifically regulating discontinuous DNA synthesis. Experimental

Procedures

Cell Labeling and Lysis Drosophila melanogaster cells (Schneider’s line 2) were maintained at 25°C. as previously described (Blumenthal and Clark, 1977). For pulse-labeling, cells in log phase were grown for 1 day in sealed plastic petri dishes, the medium was removed by aspiration and the cells were labeled for 30-120 set with fresh culture medium containing 200 pCi/ml of 3H-thymidine (AmershamSearle; spec. act. 46-50 Ci/mmole). Labeling was terminated by rapid aspiration of the medium and addition of lysis solution [I% Sarkosyl, 100 mM EDTA, 25 mM EGTA, 10 mM Tris-HCI (pH 7.9) and 100 pg/ml proteinase K (E. Merck)] at 37°C. The dishes were sealed and incubated at 45°C for 3 hr. Cell lysates were then gently scraped from the dishes, transferred to plastic culture tubes, adjusted to 0.1 N NaOH and stored at 4°C until they were fractionated on velocity gradients. When samples were to be analyzed by electrophoresis in tube gels, labeling medium was removed and replaced with 0.2 N NaOH, 0.5% Sarkosyl and 10 mM EDTA at 0°C. Cell labeling and lysis conditions have been discussed in a previous report (Blumenthal and Clark, 1977). Fractionation of Lysates The lysates from each pulse-labeling experiment were fractionated on 6 or 12 replicate 5-20% alkaline sucrose velocity gradients in an SW40 rotor, as previously described (Blumenthal and Clark, 1977), except that the EDTA concentration in the gradients was increased to 25 mM. Centrifugation was for 43 hr (w2t = 256,000 x 10’ radz/sec) at 4°C. Recovery of 3H-labeled DNA from the gradients was about 80%. Part of each fraction (150 ~1) from one gradient was assayed for 3H-thymidine-labeled, acid-precipitable molecules. We pooled fractions from three regions, designated A, B and C (see Figure I), of the 6 or 12 replicate gradients. The divisions of the gradients sometimes varied by one fraction. The pooled fractions were neutralized by dropwise addition of glacial acetic acid and then adjusted to 0.3 M sodium acetate. The C region fractions were further diluted with an equal volume of 0.3 M sodium acetate to reduce the sucrose concentration. E. coli B tRNA (2-10 pg/ml; Schwa&Mann) was added to the B and C region pools, and all pooled fractions were precipitated overnight at -20°C with 2 vol of 95% ethanol. The precipitates were collected by centrifugation at 9000 rpm and 4°C for 15-30 min in a Sorvall HB-4 rotor. They were resuspended in 70% ethanol, stored at -20°C overnight and collected by centrifugation. This was repeated once, and the precipitates were then dried under vacuum and stored at -20°C. Recovery of 3H-labeled DNA in the redissolved precipitates was 70-95%. The DNA collected by this procedure was reanalyzed on alkaline velocity gradients and gave distributions similar to those in the initial gradients. Extraction of the lysates with phenol or chloroform-isoamyl alcohol resulted in preferential loss of small DNA molecules (~3s). Sample and Marker Preparation Dried ethanol precipitates, subsequently run on slab gels, were redissolved in 50 ~1 of 10 mM Tris-HCI (pH 8) at 80-90°C for 2-5 min and cooled to 0°C. Deionized formamide (100 PI), glycerol (5 PI), 0.025% bromophenol blue (2 ~1) and 0.5% xylene cyanol FF (1

Drosophila 189

Replication

Intermediates

~1) were added to each sample. Samples were underlayered in wells filled with the running buffer. The procedure was identical for tube gels, except that 90 /LI of Tris-HCI (pH 8.0), 5 mM EDTA, IO ~1 glycerol and 200 ~1 deionized formamide were used. The samples were applied to the tops of tube gels, which had been blotted dry. Formamide (66%, v/v) containing 0.02 M sodium phosphate buffer (pH 7.5) was layered over the sample, and the running buffer was then added. Markers were prepared from 3H- or 14C-thymidine-labeled SV40 DNA (supplied by Dr. Jack Griffith, Stanford University, or by Dr. Jon Williams of this laboratory). Form I SV40 DNA was purified and digested with Hae Ill, Hind Ill (New England Biolabs) or Hind II and Ill (Miles Laboratories) restriction endonucleases. j4C-thymidine-labeled Hind II and Ill fragments of SV40 DNA were purchased from Bethesda Research Labs. This material contained primarily the Hind II fragments. Purified, 3H-labeled +X174 RF DNA was supplied by Dr. Shlomo Eisenberg (Stanford University) and was digested with the Hae Ill enzyme. Markers were heated to 80°C for 2-5 min before electrophoresis, added to formamide and run in wells adjacent to the samples. Gel Electrophoresis Polyacrylamide gels (5 or 8%) containing 98% deionized formamide (Matheson, Coleman and Bell) were prepared as described by Maniatis, Jeffrey and van deSande (1975). Slab gels (10 cm wide, 14 cm long and 1.5 mm thick) were allowed to polymerize at room temperature overnight and then prerun for 1 hr at 100 V and 20°C. After the samples were applied, the gels were run at 100 V until the bromophenol blue marker migrated 8-9 cm from the origin. Markers were run in wells adjacent to the samples and provided linear calibration curves for determination of molecular lengths. Tube gels (10 or 15 cm long with a 0.6 cm internal diameter) were prerun at 2 mamp per tube for 1 hr at 20°C. After the samples were applied, 5% polyacrylamide gels were run at 4 mamp per tube and 8% polyacrylamide gels at 2 mamp per tube until the bromophenol blue migrated about 85% of the gel length. Markers were run at the same time in separate tubes. Larger marker fragments gave sharp peaks, but shorter fragments were poorly resolved from the background level of radioactivity. This problem was not encountered in slab gels because of the increased resolution of fluorography. This difficulty and variations in length of tube gels and in slice thickness gave unreliable marker calibration of molecular lengths. The running buffer for all gels was 0.02 M sodium phosphate (pH 7.5). The running buffer and gels were maintained at 20°C during electrophoresis. Gel Analysis Slab gels were fixed in 10% TCA after electrophoresis and prepared for fluorography (Bonner and Laskey, 1974) using PPO dissolved in dimethylsulfoxide. Dried gels were exposed to Kodak X-Omat R film for 3 days to 4 weeks at -70°C. The film was usually preexposed to a relative optical density of 0.3 (Laskey and Mills, 1975). Fluorograms were scanned with a Joyce-Lobe1 microdensitometer to obtain optical density profiles. Tube gels were partially frozen on dry ice and sliced in 1 .I mm sections with a Hoeffer gel slicer (Hoeffer Instruments). The slices were placed in glass scintillation vials and digested with 7 ml of 3% Protosol (New England Nuclear) in Omnifluor (New England Nuclear; 15.14 g/gal of toluene) for 3 days at 37°C or overnight at 55°C. The samples were then cooled to -20°C for 3 hr, and the tritium content was determined, with 36% efficiency, in an LKB liquid scintillation counter at 10°C. All averages are given t 1 standard error of the mean. Acknowledgments This work was supported opment Administration.

by the U.S. Energy

Received

revised

April

30, 1977;

June

Research

13, 1977

and Devel-

References Blumenthal, 26.

A. B. and Clark,

Bonner, 88.

W. M. and Laskey,

Carroll,

D. and Brown,

E. J. (1977). R.A.

(1974).

D. D. (1976).

Exp.

Eur. J. Biochem.46,

Garrard,

W. T. and Todd,

R. D. (1977).

J. R. and Clarkson,

Kasamatsu, 1945-1949. Kowalski, 615. Kuebbing, 55-66.

Proc.

Cold

Spring

J. Mol.

R. (1976).

Nat.

in press.

Eur. J. Biochem.

Nat. Acad.

W. P. (1976).

D., Diaz, A. T. and Werner,

Proc.

J. Biol. Chem.,

H. (1974).

H. and Wu, M. (1976). J. and Cheevers,

P. (1976).

J. M. (1975).

Huberman, J. A. and Horowitz, Symp. Quant. Biol. 38, 233-238.

83-

Cell 7, 467-475.

Compton, J. L., Bellard, M. and Chambon, Acad. Sci. USA 73, 4382-4386. Gautschi, 403-412.

Cell Res. 105, 15

50,

Harbor

Sci. USA 73, Biol.

104, 603-

J. Mol.

Biol. 708,

Laird, C. D., Chooi, W. Y., Cohen, E. H., Dickson, f., Hutchinson, N. and Turner, S. H. (1974). Cold Spring Harbor Symp. Quant. Biol. 38, 311-327. Laskey, 341.

R. A. and

Mills,

A. D. (1975).

Lindahl,

T. (1976).

Nature

259, 64-66.

Maniatis, T., Jeffrey, 14, 3787-3794.

Eur.

A. and van deSande,

J. Biochem.

H. (1975).

Okazaki, R., Okazaki, T., Sakabe, K., Sugimoto, (1968). Proc. Nat. Acad. Sci. USA 59, 598-605.

Biochemistry

K. and Sugino,

Scott, J. F., Eisenberg, S., Bertsch, L. L. and Kornberg, Proc. Nat. Acad. Sci. USA 74, 193-197. Tseng,

B. Y. and Goulian,

Tye, B-K., B. (1977). Weintraub,

M. (1975).

J. Mol.

A. and Alberts,

A.

A. (1977).

Biol. 99, 317-337.

Nyman. P-O., Lehman, I. R., Hochhauser, Proc. Nat. Acad. Sci. USA 74, 154-157. H., Worcel,

56, 335

B. (1976).

S. and Weiss, Cell 9, 409-417.