- Email: [email protected]
polypurine DNA
Cell, Vol. 5, 173-181,
June
1975,
Copyright
0 1975
by MIT
Polypyrimidine Segments in Drosophila DNA: I. Detection of a Cryptic Satellite Polypyrimidine/Polypurine DNA H. C. Birnboim Biology and Health Physics Division Atomic Energy of Canada Limited Chalk River, Ontario KOJ 1JO Ft. Sederoff Department of Biological Sciences Columbia University New York, New York 10027
Summary Very long runs of pyrimidine nucleotides (polypyrimidines), previously detected in DNA from Drosophila melanogaster, have now been localized to a “cryptic” satellite. These polypyrimidines have an average length of 750 nucleotides and account for about 3% of the thymine residues in total DNA. The buoyant density of the DNA component which contains the polypyrimidines was detected by centrifuging native DNA to equilibrium in a CsCl gradient, and then assaying each fraction for its content of polypyrimidines. A peak was detected at a density of about 1.707 gm/cm3, distinctly heavier than the main band of DNA (1.702 gm/ cm3). The buoyant density of polypyrimidine-containing molecules was little affected by differences in the molecular weight of the starting DNA in the range 1 OS-107 daltons (single-stranded). Thus polypyrimidines (and their complementary polypurines) appear to form all or part of a “cryptic” satellite. Polypyrimidines have been isolated and characterized with respect to composition and buoyant density. Direct nucleoside analysis of unlabeled material indicated 34.5% deoxycytidine, 65.5% thymidine. Their banding position in neutral and alkaline CsCl gradients was consistent with a singlestranded DNA polymer of this composition. Introduction As in the case of many other organisms, DNA from Drosophila melanogaster forms discrete satellite bands when centrifuged to equilibrium in CsCl density gradients (Travaglini, Petrovic, and Schultz, 1968; Laird and McCarthy, 1969; Gall, Cohen, and Polan, 1971; Travaglini, Petrovic, and Schultz, 1972). Additional satellite components, obscured because their density is very close to main band DNA, have been resolved by centrifugation in antibiotic-CsCI gradients (Peacock et al., 1973). We have been studying a component of D. melanogaster DNA whose annealing properties suggested it was part of a highly repititious portion of the genome, but its composition was unlike that of any of the known satellites: it was composed of long uninter-
melanogaster Containing
rupted runs of pyrimidine nucleotides, averaging 750 residues in length (Birnboim, Straus, and Sederoff, 1975). Because of their highly repetitious nature, it seemed possible that polypyrimidines were part of a satellite component whose composition had not previously been determined. In this communication we present evidence which supports this hypothesis; polypyrimidine-containing molecules were found to band in CsCl gradients as a cryptic satellite whose density was slightly higher than main-band DNA. Since pyrimidine polynucleotides of this length have not been detected previously in DNA, we have studied some of their physical and chemical properties. Earlier it was shown that 3H-thymidine-labeled polypyrimidines can anneal very rapidly to homologous DNA, but not to poly(A) (Birnboim et al., 1975). This indicated a simple sequence containing both thymidine and deoxycytidine residues. In the present study we have isolated 3H-thymidinelabeled as well as unlabeled polypyrimidines and have determined their UV spectrum, nucleoside composition, and buoyant density in CsCl gradients. These experiments lend further support to our earlier suggestion that they are polydeoxynucleotides composed only of pyrimidine residues. Results Isolation and Characterization of Unlabeled Polypyrimidines To permit direct nucleotide analysis and for use as a template for RNA polymerase in other experiments, unlabeled polypyrimidines were isolated from 10 mg of DNA from D. melanogaster cultured cells. Acid-treated DNA was chromatographed on Sephadex G-50 as described in Experimental Procedures and a distinct A254 peak was seen at the excluded volume (Figure la). The peak (polypyrimidines) was concentrated and dissolved in 0.03 N sodium acetate (pH 5.0) for spectral analysis (Figure 1 b). The observed spectrum is consistent with a polydeoxynucleotide containing only deoxycytidine and deoxythymidine (Cape, 1967). For direct analysis of composition, polypyrimidines were enzymatically hydrolyzed and the nucleoside products were separated chromatographically (Figure 1 c). The ratio of nucleosides was (dC)34.8%:(dT)65.2%; no purine nucleosides were detected. Buoyant Density of Isolated Polypyrimidines Isolated 3H-polypyrimidines were of sufficient size to permit banding in CsCl (Figure 2). Their buoyant density is estimated to be 1.75 in CsCl (pH 8.0) and 1.76 in alkaline CsCI. The density of an unusual polymer such as Drosophila polypyrimidines may not
Cell 174
be predicted accurately from the relationship between composition and density of single-stranded polymers in neutral CsCl derived by Riva et al. (1969). For example, the observed and predicted values for poly(dC) differ significantly (Szybalski and Szybalski, 1971). However, the predicted value for polypyrimidines [655%(dT); 34.5%(dC)] . is 1.751, close to the observed value. Under alkaline conditions, a weighted average using the densities of poly(dT) (1.771) and poly(dC) (1.722) (Szybalski and Szybalski, 1971) gives a value of 1.754, again close to the observed figure. Despite the uncertainty involved in both types of calculation, we feel that this experiment lends further support to our proposal that bona fide polypyrimidines may be isolated from labeled and unlabeled Drosophila DNA.
I
I
I
I
I I. 80
18C
I
I. 15
3
l4C
125
lO[
I5
6t
r cc x k
I. 70
20
180
190
b
I25
100
60
20 10
20 Fraction
Figure 1. Analysis of Polypyrimidines melanogaster DNA
from
Unlabeled
Drosophila
(a) Chromatography on Sephadex G-50 of acid-treated DNA. Polypyrimidines appear as a peak at the excluded volume (V,). (b) UV absorbance spectrum of polypyrimidines from (a). The solvent was 0.03 M sodium acetate (pH 5.0). (c) Nucleoside analysis of polypyrimidines, which had been digested with venom phosphodiesterase, pancreatic deoxyribonuclease and alkaline phosphatase. Chromatography on Aminex A-6 is described in Experimental Procedures. Minor peaks 1 and 2 correspond to nonnucleotide material (enzyme, salts) in the digest. Peak 3 is due to a baseline shift.
30
90
50
number
Figure 2. Banding in CsCl Gradients of Isolated XH-Polypyrimidines JH-thymidine-labeled polypyrimidines, isolated as described in Experimental Procedures, were centrifuged to equilibrium in (a) neutral CsCl (0.015 M Tris, pH 8.0, 3 mM EDTA, 0.5% Sarkosyl) or (b) alkaline CsCl (0.1 N NaOH). The initial density of the CsCl solutions was 1.75 g/cm3. Samples (6.5 ml) were centrifuged for 72 hr, 2O”C, in a Beckman type 65 fixed angle rotor at 40,000 rpm. The density of selected fractions, as determined from their refractive index, is shown in the inset. TCA-precipitable radioactivity in 3H-polypyrimidines (-O-) and in E. coli marker W-DNA (-C-) is shown.
Polypyrimidines 175
(TOP)
in a Cryptic
Satellite
10
20
30
40
Sedimentation
Analysis
Fraction Figure
3. Alkaline
Sucrose
Gradient
IO
30
20
40
Number
of 3H-Polypyrimidines
(a) After hydrolysis of 3H-thymidine-labeled DNA with formic acid/diphenylamine, resistant pyrimidine tracts were precipitated either with ethanol (standard procedure, -O-) or with acid (ARAP procedure, -c-). The precipitated tracts were analyzed by alkaline sucrose gradient sedimentation. (b) The ARAP procedure was used to assay polypyrimidines in rapidly-reassociating DNA (-0-) and slowly-reassociating DNA (-a-) which had been separated using hydroxyapatite. Gradients 40 rotor
were prepared using 5% and 20% at 38,000 rpm for 17 hr at 20°C.
(w/v)
Assay of Polypyrimidine Segments as Acid-Resistant Acid-Precipitable (ARAP) Radioactivity In previous experiments long pyrimidine acid-hydrolyzed Drosophila DNA were from shorter tracts by ethanol precipitation kaline sucrose gradient sedimentation Straus, and Sederoff, 1975). To facilitate s/s of multiple samples of DNA for their polypyrimidines, a more rapid technique
sucrose
and
tracts in separated and al(Birnboim, the analycontent of has been
0.9 N NaCI-0.1
N NaCH.
Samples
were
centrifuged
in a Beckman
SW
developed. This method is based upon selective precipitation with HCI of long pyrimidine tracts after treatment of DNA with formic acid/diphenylamine. The ARAP method and the standard method (which involves ether extraction and ethanol precipitation) were compared by checking the size of the precipitated polynucleotides on alkaline sucrose gradients (Figure 3a).
Cell 176
The results of such an experiment show that both small and large oligonucleotides were precipitated by the standard method but only large nucleotide fragments were precipitated by the ARAP method (Figure 3a). An additional advantage of the ARAP method is an improved recovery of polypyrimidines, some of which may be lost at the ether extraction step in the standard method (Birnboim et al., 1975). Another demonstration of the ARAP method is shown in Figure 3b. Rapidly reassociating Drosophila DNA was isolated by hydroxyapatite chromatography (Botchan et al., 1971; Kram, Botchan, and Hearst, 1972). Polypyrimidines have a very low Cot, of reassociation when annealed to D. melanogaster DNA (Birnboim et al., 1975); therefore, we expect polypyrimidines to be bound to hydroxyapatite as part of the rapidly reassociated fraction. Both bound and unbound fractions were subjected to ARAP treatment, and the final precipitates were analyzed by alkaline sucrose gradient centrifugation (Figure 3b). As expected, polypyrimidines were found exclusively as part of rapidly reassociating DNA which bound to hydroxyapatite. This experiment also demonstrates the utility of the ARAP procedure as a direct method for quantitating polypyrimidines, since unbound DNA (which lacked polypyrimidines) yielded very little ARAP radioactivity. In another experiment, there was essentially complete recovery of purified 3H polypyrimidines when mixed with unlabeled DNA and carried through the ARAP procedure (data not shown). Because it gave complete recovery of polypyrimidines and little contamination with short pyrimidine tracts, ARAP radioactivity was used as a direct assay for polypyrimidines in subsequent experiments. Distribution of Polypyrimidine-Containing DNA Molecules in CsCl Gradients From the data shown in Figure 3a and other experiments (Birnboim’et al., 1975) we have estimated that the weight average molecular weight of polypyrimidines is 2.48 x 105 daltons. If polypyrimidines were tightly clustered, that is, in adjacent regions in either a cis or tram configuration on duplex DNA molecules, then these DNA molecules could have a buoyant density different from “main-band” DNA and might be separable on a CsCl gradient. If widely distributed throughout the genome, polypyrimidines would be on separate DNA molecules and would likely have little effect on the buoyant density of such molecules provided their size was fairly large (106-l 07 daltons). The buoyant density distribution in CsCl of total DNA (weight average molecular weight in alkaline sucrose, Mwaik = 14 X 106 daltons) and the distribution of polypyrimidine-containing DNA molecules were determined by sedimentation to equilibrium in CsCl (Figure 4). Frac-
B I
20
30 Fraction
Figure 4. Banding midine-Containing
in a CsCl Gradient Molecules
40
6
I
50
number of Total
DNA and of Polypyri-
SH-thymidine-labeled native DNA (&ark = 1.39 x 107 daltons) was centrifuged to equilibrium in a CsCl gradient (see Experimental Procedures), and a portion of each gradient fraction was tested for TCA-precipitable radioactivity (-) and ARAP radioactivity ). The density marker is native E. coli ‘AC(polypyrimidines) (-DNA.
tions were collected and assayed for TCA-precipitable radioactivity (representing total DNA) and ARAP radioactivity (representing polypyrimidine-containing molecules). The major fraction of polypyrimidine-containing molecules are seen to band a position (p g 1.707) which is distinctly heavier than main-band DNA (p = 1.702) and lighter than the E. coli DNA marker (p = 1.710). This experiment indicates that most polypyrimidines are tightly clustered within the genome and behave as a cryptic satellite. Similar experiments were carried out using DNA samples of lower molecular weight (Figure 5). Sheared DNA (MWalk = 1.47 x 106 daltons) and sonicated DNA @Latk = 3.23 x 105 daltons) were centrifuged to equilibrium in CsCl and the positions of total DNA and polypyrimidine-containing molecules were determined. In both cases polypyrimidine-containing molecules banded at a density slightly greater than main-band DNA, although in the case of sonicated DNA the zones were quite broad. These data provide strong evidence that most polypyrimidine segments are clustered-that is, many such segments are grouped together on a single DNA strand.
Polypyrimidines 177
in a Cryptic
Satellite
1 (b)
I.710
(a)
I
I
30
40
50
(TOP--I
20 FRACTION
Figure
5. Banding
in CsCl
-
Gradients
of Native
Sheared
30
DNA and of Polypyrimidine-Containing
Lightly sheared DNA (M,alr = 1.47 x 106 daltons) (a) and sonicated in CsCl gradients as described in Experimental Procedures. Each -o-) and for radioactivity in polypyrimidines (ARAP, -0-). The in alkaline sucrose gradients.
Use of Limited Acid Hydrolysis to Estimate the Extent of Clustering of Polypyrimidines We have developed another approach, different from CsCl banding, to assess the extent of clustering of polypyrimidines in the genome. It is based upon an indirect method for determining the number of purine residues in single-stranded, polypyrimidine-containing molecules. The purine content of a DNA molecule may be estimated from the rate of breakage of the molecule by limited acid treatment, since purine residues are cleaved much more readily than pyrimidine residues, and apurinic sites are broken in alkali. If polypyrimidines were clustered along one strand of DNA, then that strand would have few purines and would be relatively resistant to acid hydrolysis, as compared to “average” DNA molecules which contain about 50% purines. If polypyrimidines were widely interspersed with regions of average composition, then strands to which they were attached would be nearly as sensitive as average DNA molecules. For the experiment, we
40
50
(TOP--)6
NUMBER Molecules
DNA (M Week = 3.23 x 105 daltons) (b) were centrifuged to equilibrium gradient fraction was tested for total radioactivity (TCA-precipitable, inset shows the sedimentation profiles of the two DNA preparations
treated high molecular weight DNA with dilute acid and compared by alkaline sucrose gradient sedimentation the extent to which total DNA and polypyrimidine-containing DNA molecules were degraded (Figure 6). The conditions for limited acid hydrolysis were determined empirically in a preliminary experiment. The profiles of total DNA and polypyrimidine-containing molecules after limited acid treatment clearly indicates that the latter are much more resistant to breakdown. For a more quantitative interpretation, the weight average molecular weights were determined as described in Experimental Procedures. [Although number average molecular weight should be used to estimate strand breaks, we have used weight average molecular weight because number average is strongly biased by small fragments near the top of the gradient. Our conclusion, that there are relatively few purines on polypyrimidine-containing molecules, is not altered by this treatment of the data. Similar treatment of sucrose gradients using weight average molecular weight has been used by Regan, Setlow, and Ley
Cell 178
S 20,
w
8
16
24
32
I
I
I
I
8
6
(1971).] The size of total DNA was initially 13,100 nucleotides long, and this was reduced to 2130 nucleotides after treatment. Polypyrimidine-containing molecules (identified as ARAP radioactivity) were reduced in size from 9830 to 7430 nucleotides. These values indicate that there are only about 8% as many purines in polypyrimidine-containing molecules as there are in “average” DNA (Table 1). Thus it appears that many polypyrimidine segments are grouped together on a single strand of DNA.
Discussion The data presented here and in an earlier report (Birnboim et al., 1975) indicate that an unusual polynucleotide segment can be isolated from Drosophila melanogaster DNA. This segment contains, on the average, 750 pyrimidine nucleotides (34.5% deoxycytidine and 65.5% thymidine) and accounts for about 3% of the thymidine residues of total DNA. By several criteria, including (a) buoyant density in neutral and alkaline CsCl gradients, (b) UV spectral analysis and nucleoside composition, (c) resistance to formic acid/diphenylamine, and (d) ability to anneal at low Cot to homologous DNA, it appears to be an authentic DNA polymer. The large size and unusual composition of these polypyrimidines have provided the basis for an ARAP assay (acid resistant-acid precipitable method), suitable for the analysis of multiple samples from CsCl and alkaline sucrose gradients. The rapid rate of annealing of polypyrimidines with DNA led us to suspect that they had a simple sequence and therefore might be organized as a satellite component. No similar long pyrimidine tracts had been seen previously by Peacock et al. (1973) in the satellites they had examined. We found that polypyrimidine-containing molecules, identified by the ARAP method, were localized principally to a cryptic satellite which bands in CsCl at about 1.707, 5 mg/cmx heavier than main-band DNA. A satellite having a density of 1.705 has recently been identified by Ag+-CsS04 gradient sedimentation and found to direct the synthesis of polypurine complementary RNA (Endow, Polan, and Gall, personal communication). It may be the same as the 1.705 satellite described by Peacock et al.
8
6
Figure 6. Sedimentation in Alkaline Sucrose Gradients DNA and of DNA Treated Briefly with Acid
4
8
Fraction
12
16
20 211
number
of Untreated
The sedimentation profile of total DNA (-0-) and of polypyrimidine-containing molecules (-•-), as assayed by the ARAP method, was determined. In the top figure, untreated DNA was sedimented. In the lower figure, DNA was treated briefly with dilute HCI to introduce a small number of breaks at purine sites before sedimentation. Polypyrimidine-containing molecules changed very little in size, indicating that polypyrimidine segments are tightly clustered (Table 1). Centrifugation was for 16 hr, 24,000 rpm at 20°C.
Polypyrimidines 179
Table
in a Cryptic
1. Calculation
Satellite
of the Number
of Purine
Nucleotides
in Polypyrimidine-Containing Total
Initial Size (nucleotides) Size after
limited
hydrolysis
Breaks
per molecule
a--I ( b Breaks
) per 1000
a
Molecules PolypyrimidineContaining Molecules
DNA
13,100
9830
2,130
7430
5.15
0.32
0.39
0.033
nucleotides
cxE ( Purines
DNA
1
per break 1,282
(“> 2c Purines
per 1000
nucleotides
Cd x e)
500
42
(assumed)
(calculated)
These calculations are based upon the experiment of Figure 6. The sizes of the DNA samples as given in the Table are derived from the weight average molecular weights; although number average molecular weights should theoretically be used, the practical difficulties in using this value have been noted previously (Regan et al., 1971). Isolated polypyrimidines are 750 nucleotides long, as calculated from the weight average molecular weight of their sedimentation profiles.
(1973), although their pyrimidine tract analysis showed no polypyrimidine component. Selective losses of polypyrimidines after formic acid-diphenylamine treatment have been seen previously(Birnboim et al., 1975) and may be responsible for the differences. Several possible models by which polypyrimidines (and their complementary polypurines) may be arranged in the genome are depicted in Figure 7. In model (a), polypyrimidine/polypurine segments are widely scattered throughout the genome, so that not more than one segment is associated with any DNA molecule as normally isolated. In such a case, where the length of the DNA molecule is large relative to the polypyrimidine/polypurine segment, there would be no apparent buoyant density shift. This model is ruled out because a buoyant density shift was observed (Figures 4 and 5). In model (c), the segments are clustered but tandemly inverted in such a way that polypyrimidines and polypurines alternate along each strand. In this model, each strand contains 50% purines and therefore should be as sensitive to partial acid hydrolysis as “average” DNA. This was not the case (Figure 6), making this model unlikely. In model (b), polypyrimidines are completely contiguous-that is, one of the two DNA strands is composed only of pyrimidine residues. If this were the case, polypyrimidines isolated after formic acid-diphenylamine treatment should be as long as the total DNA strand provided nonspecific breaks (at pyrimidine sites) do not occur. A small amount of nonspecific breakage has been observed under these conditions of acid
P
DNA polypyrimidine Polypurine Figure 7. Models in DNA (a) (b) (c) (d)
of the Possible
strand
_
segment
-II
segment
I
Arrangement
of Polypyrimidines
Widely interspersed. Contiguous. Tandemly inverted. Tightly clustered.
Models(b), (c), ments (Figure DNA band as ment (Figure strands which
and(d) are consistent with the CsCl banding experi4 and 5) which show that polypyrimidines in native a “cryptic” satellite. The limited hydrolysis experi6) suggests that polypyrimidines are part of DNA contain relatively few purine residues (model d).
treatment (Birnboim et al., 1975). However, the very mild acid treatment used in the experiment of Figure 6 led to breakage of only about 1 in 1000 purine sites, so no detectable breaks at pyrimidine sites would be anticipated. Figure 6 and Table 1 show
Cl?ll 180
that some change in size of polypyrimidine-containing molecules was detected, indicating that some purines were present. Model(b) is therefore unlikely to be valid. The data are consistent with model (d), in which polypyrimidines are tightly clustered along one strand, separated by, on the average, 33 purines. In an accompanying report (Sederoff, Lowenstein, and Birnboim, 1975), we present evidence concerning the chromosomal location and nucleotide sequence of the polypyrimidine/polypurine segments of D. melanogaster DNA. This component may prove to be particularly useful for evolutionary studies of satellites because a relatively low frequency of purine substitutions in the polypyrimidine regions should be detectable by the methods we have described.
ed by centrifugation and dissolved in 2 ml 0.1 M sodium pyrophosphate at 60°C. For counting, 0.4 ml of 1 N HCI and 10 ml PCS scintillation fluid (Searle Instrumentation) were added. In some experiments, the final pellet was dissolved in 0.2 ml of 0.1 M sodium pyrophosphate and layered on an alkaline sucrose gradient.
Experimental
CsCl Equilibrium Density Gradient Cenfrifugation ,H-thymidine-labeled DNA and polypyrimidine samples, together with an internal E. coli I4C-DNA marker, were centrifuged to equilibrium in CsCl gradients containing either 15 mM Tris-HCI, 3 mM EDTA, 0.5% Sarkosyl (pH 8.0) or 0.1 N NaOH. 10 ml gradients were centrifuged in polyallamer tubes in a Beckman type 65 rotor at 35,000 rpm for at least 60 hr at 20°C, unless otherwise indicated. Fractions were collected dropwise from the bottom, and acidprecipitable counts were determined.
Procedures
Conditions of Cell Culture, Labeling with ‘H-Thymidine, and Preparation of DNA The detailed conditions for growth and labeling of Drosophila melanogaster cells (Schneider’s line 2) and the preparation of SHthymidine-labeled DNA has been described elsewhere (Birnboim et al., 1975). In some experiments, DNA was prepared by the ureahydroxyapatite procedure described by Britten, Graham, and Neufeld (1974). Isolation of Unlabeled Polypyrimidines Ten mg of DNA from cultured ceils was partially hydrolyzed with HCI (0.1 N, lOO”C, 30 min), and the fraction which precipitated on standing at 0°C was treated with 67% formic acid, 2% diphenylamine at 30°C for 40 hr. Polypyrimidines were isolated from the hydrolysate using Sephadex G-50 chromatography as described elsewhere (Birnboim et al., 1975). Base Composition Analysis and UV Spectrum Unlabeled polypyrimidines were dissolved in 0.1 M Tris-HCI (pH 8.0), 1.5 mM MgC12, and hydrolyzed to the nucleoside level using a mixture of venom phosphodiesterase, pancreatic DNAase, and alkaline phosphatase. About 20 nmoles of nucleosides were charged onto a column of Aminex A-6 (0.9 x 25 cm) and eluted with 0.3 M ammonium acetate (pH 5.7) (Singhal, 1972). Nucleosides were monitored with a Chromatronix Model 200 UV detector, and their molar ratio was calculated by reference to standard nucleoside solutions. The UV absorbance spectrum of polypyrimidines was determined using a Beckman Acta V spectrophotometer. Assay for Polypyrimidines as Acid-Resistant Acid-Precipitable (ARAP) Radioactivity 3H-thymidine-labeled DNA and 100 pg carrier DNA were dissolved in 0.25 ml HzO, and 0.75 ml of 90% formic acid-2.7% diphenylamine was added. Hydrolysis was allowed to proceed in the dark at 30°C for 18-20 hr. Long pyrimidine tracts were selectively precipitated from the hydrolysis mixture as follows, To each sample, 0.1 ml of carrier DNA (1 mg/ml), 0.1 ml of 0.1 M sodium pyrophosphate, 0.1 ml of 3 N HCI, and 3 ml of cold ethanol were added in that order. After 30 min at O”C, the precipitate which formed was collected by centrifugation at 30,000 x g for 15 min at 0°C. The pellet was redissolved in 0.8 ml of 0.1 M sodium pyrophosphate, heated at 60°C for 5 min, and reprecipitated at 0°C by the addition of 0.4 ml of 3 N HCI. After 20 min, the sample was centrifuged and the pellet was again redissolved in 0.8 ml of 0.1 M sodium pyrophosphate as before. Following the addition of 0.2 ml of 3 N HCI and a 20 min period of standing at O”C, the final precipitate was collect-
Alkaline Sucrose Gradients for Estimation of DNA Size AH-thymidine-labeled DNA or polypyrimidines were dissolved in 0.2 ml of 0.1 M sodium pyrophosphate and layered over 13.2 ml linear sucrose gradients (5 to 20% w/v in 0.9 N NaCI, 0.1 N NaOH). Gradients were centrifuged in a Beckman SW40 rotor at 20°C for appropriate times and fractionated as described elsewhere (Birnboim, 1972). Sedimentation values were calculated by a procedure similar to that used by McEwen (1967); a program for this purpose using a Hewlett-Packard 9820 A calculator was developed by R. Holford (Atomic Energy of Canada Limited). S values were converted to molecular weights using the relationship of Studier (1265). Weight average molecular weights were calculated as M, = Z qM,/I: ci, where c is radioactivity in cpm and M is calculated molecular weight. [Copies of this report are available from Atomic Energy of Canada Limited, Chalk River: No. AECL-5022.1
Acknowledgments We thank Renate Clynes and J. M. Ostrom for excellent technical assistance. This work was supported in part by an NIH grant. We are grateful to R. Holford of the Health Physics Branch (Atomic Energy of Canada Limited) for developing a computer program for calculating S values and weight average molecular weights. Received
February
6, 1975
References Birnboim,
H. C. (1972).
Biochemistry
Birnboim, H. C., Straus, istry 14, 1643.
17, 4588.
N. A., and Sederoff,
R. R. (1975).
Botchan, M., Kram, R., Schmid, C. W., and Hearst, Proc. Nat. Acad. Sci. USA 68, 1125.
Biochem-
J. E. (1971).
Britten, R. J., Graham, D. E., and Neufeld, B. R. (1974). In Methods in Enzymology, 29, L. Grossman and K. Moldave, eds. (New York: Academic Press), p. 363. Cape, R. (1967). Canada.
Ph.D. Thesis,
McGill
Gall, J. G., Cohen, 33, 319.
E. H., and
Kram, 103.
R., Botchan,
M., and Hearst,
Laird,
C., and
McEwen,
McCarthy,
C. R. (1967).
Polan,
B. (1969). Anal.
University,
Montreal,
M. L. (1971). J. E. (1972). Genetics
Biochem.
Quebec,
Chromosoma J. Mol. Biol. 64,
63, 865.
20, 114.
Peacock, W. J., Brutlag, D., Goldring, E., Appels, R., C. W., and Lindsley, D. L. (1973). Cold Spring Harbor Quant. Biol. 38, 405. Regan, J. D., Setlow, Sci. USA 68, 708.
R. B., and Ley, R. D. (1971).
Riva, S., Barrai, I,, Cavalli-Sforza, Mol. Biol. 45, 367.
L., and
Proc.
Falaschi,
Hinton, Symp.
Nat. Acad. A. (1969).
J.
Polypyrimidines 181
Sederoff, 5, 183.
in a Cryptic
R.,
Lowenstein,
Satellite
L.,
and
Singhal,
R. P. (1972).
Arch.
Studier,
F. W. (1965).
J. Mol. Biol.
Birnboim,
Biochem.
H. C. (1975).
Biophys. II,
Cell
152, 800.
373.
Szybalski, W., and Szybalski, E. H. (1971). In Procedures in Nucleic Acid Research, 2, G. L. Cantoni and D. R. Davies, eds. (New York: Harper and Row), p. 311. Travaglini, 136a.
E., Petrovic,
Travaglini, 419.
E. C., Petrovic,
J., and Schultz, J., and Schultz,
J. (1968). J. (1972).
J. Cell Biol. 39, Genetics
72,
June
1975,
Copyright
0 1975
by MIT
Polypyrimidine Segments in Drosophila DNA: I. Detection of a Cryptic Satellite Polypyrimidine/Polypurine DNA H. C. Birnboim Biology and Health Physics Division Atomic Energy of Canada Limited Chalk River, Ontario KOJ 1JO Ft. Sederoff Department of Biological Sciences Columbia University New York, New York 10027
Summary Very long runs of pyrimidine nucleotides (polypyrimidines), previously detected in DNA from Drosophila melanogaster, have now been localized to a “cryptic” satellite. These polypyrimidines have an average length of 750 nucleotides and account for about 3% of the thymine residues in total DNA. The buoyant density of the DNA component which contains the polypyrimidines was detected by centrifuging native DNA to equilibrium in a CsCl gradient, and then assaying each fraction for its content of polypyrimidines. A peak was detected at a density of about 1.707 gm/cm3, distinctly heavier than the main band of DNA (1.702 gm/ cm3). The buoyant density of polypyrimidine-containing molecules was little affected by differences in the molecular weight of the starting DNA in the range 1 OS-107 daltons (single-stranded). Thus polypyrimidines (and their complementary polypurines) appear to form all or part of a “cryptic” satellite. Polypyrimidines have been isolated and characterized with respect to composition and buoyant density. Direct nucleoside analysis of unlabeled material indicated 34.5% deoxycytidine, 65.5% thymidine. Their banding position in neutral and alkaline CsCl gradients was consistent with a singlestranded DNA polymer of this composition. Introduction As in the case of many other organisms, DNA from Drosophila melanogaster forms discrete satellite bands when centrifuged to equilibrium in CsCl density gradients (Travaglini, Petrovic, and Schultz, 1968; Laird and McCarthy, 1969; Gall, Cohen, and Polan, 1971; Travaglini, Petrovic, and Schultz, 1972). Additional satellite components, obscured because their density is very close to main band DNA, have been resolved by centrifugation in antibiotic-CsCI gradients (Peacock et al., 1973). We have been studying a component of D. melanogaster DNA whose annealing properties suggested it was part of a highly repititious portion of the genome, but its composition was unlike that of any of the known satellites: it was composed of long uninter-
melanogaster Containing
rupted runs of pyrimidine nucleotides, averaging 750 residues in length (Birnboim, Straus, and Sederoff, 1975). Because of their highly repetitious nature, it seemed possible that polypyrimidines were part of a satellite component whose composition had not previously been determined. In this communication we present evidence which supports this hypothesis; polypyrimidine-containing molecules were found to band in CsCl gradients as a cryptic satellite whose density was slightly higher than main-band DNA. Since pyrimidine polynucleotides of this length have not been detected previously in DNA, we have studied some of their physical and chemical properties. Earlier it was shown that 3H-thymidine-labeled polypyrimidines can anneal very rapidly to homologous DNA, but not to poly(A) (Birnboim et al., 1975). This indicated a simple sequence containing both thymidine and deoxycytidine residues. In the present study we have isolated 3H-thymidinelabeled as well as unlabeled polypyrimidines and have determined their UV spectrum, nucleoside composition, and buoyant density in CsCl gradients. These experiments lend further support to our earlier suggestion that they are polydeoxynucleotides composed only of pyrimidine residues. Results Isolation and Characterization of Unlabeled Polypyrimidines To permit direct nucleotide analysis and for use as a template for RNA polymerase in other experiments, unlabeled polypyrimidines were isolated from 10 mg of DNA from D. melanogaster cultured cells. Acid-treated DNA was chromatographed on Sephadex G-50 as described in Experimental Procedures and a distinct A254 peak was seen at the excluded volume (Figure la). The peak (polypyrimidines) was concentrated and dissolved in 0.03 N sodium acetate (pH 5.0) for spectral analysis (Figure 1 b). The observed spectrum is consistent with a polydeoxynucleotide containing only deoxycytidine and deoxythymidine (Cape, 1967). For direct analysis of composition, polypyrimidines were enzymatically hydrolyzed and the nucleoside products were separated chromatographically (Figure 1 c). The ratio of nucleosides was (dC)34.8%:(dT)65.2%; no purine nucleosides were detected. Buoyant Density of Isolated Polypyrimidines Isolated 3H-polypyrimidines were of sufficient size to permit banding in CsCl (Figure 2). Their buoyant density is estimated to be 1.75 in CsCl (pH 8.0) and 1.76 in alkaline CsCI. The density of an unusual polymer such as Drosophila polypyrimidines may not
Cell 174
be predicted accurately from the relationship between composition and density of single-stranded polymers in neutral CsCl derived by Riva et al. (1969). For example, the observed and predicted values for poly(dC) differ significantly (Szybalski and Szybalski, 1971). However, the predicted value for polypyrimidines [655%(dT); 34.5%(dC)] . is 1.751, close to the observed value. Under alkaline conditions, a weighted average using the densities of poly(dT) (1.771) and poly(dC) (1.722) (Szybalski and Szybalski, 1971) gives a value of 1.754, again close to the observed figure. Despite the uncertainty involved in both types of calculation, we feel that this experiment lends further support to our proposal that bona fide polypyrimidines may be isolated from labeled and unlabeled Drosophila DNA.
I
I
I
I
I I. 80
18C
I
I. 15
3
l4C
125
lO[
I5
6t
r cc x k
I. 70
20
180
190
b
I25
100
60
20 10
20 Fraction
Figure 1. Analysis of Polypyrimidines melanogaster DNA
from
Unlabeled
Drosophila
(a) Chromatography on Sephadex G-50 of acid-treated DNA. Polypyrimidines appear as a peak at the excluded volume (V,). (b) UV absorbance spectrum of polypyrimidines from (a). The solvent was 0.03 M sodium acetate (pH 5.0). (c) Nucleoside analysis of polypyrimidines, which had been digested with venom phosphodiesterase, pancreatic deoxyribonuclease and alkaline phosphatase. Chromatography on Aminex A-6 is described in Experimental Procedures. Minor peaks 1 and 2 correspond to nonnucleotide material (enzyme, salts) in the digest. Peak 3 is due to a baseline shift.
30
90
50
number
Figure 2. Banding in CsCl Gradients of Isolated XH-Polypyrimidines JH-thymidine-labeled polypyrimidines, isolated as described in Experimental Procedures, were centrifuged to equilibrium in (a) neutral CsCl (0.015 M Tris, pH 8.0, 3 mM EDTA, 0.5% Sarkosyl) or (b) alkaline CsCl (0.1 N NaOH). The initial density of the CsCl solutions was 1.75 g/cm3. Samples (6.5 ml) were centrifuged for 72 hr, 2O”C, in a Beckman type 65 fixed angle rotor at 40,000 rpm. The density of selected fractions, as determined from their refractive index, is shown in the inset. TCA-precipitable radioactivity in 3H-polypyrimidines (-O-) and in E. coli marker W-DNA (-C-) is shown.
Polypyrimidines 175
(TOP)
in a Cryptic
Satellite
10
20
30
40
Sedimentation
Analysis
Fraction Figure
3. Alkaline
Sucrose
Gradient
IO
30
20
40
Number
of 3H-Polypyrimidines
(a) After hydrolysis of 3H-thymidine-labeled DNA with formic acid/diphenylamine, resistant pyrimidine tracts were precipitated either with ethanol (standard procedure, -O-) or with acid (ARAP procedure, -c-). The precipitated tracts were analyzed by alkaline sucrose gradient sedimentation. (b) The ARAP procedure was used to assay polypyrimidines in rapidly-reassociating DNA (-0-) and slowly-reassociating DNA (-a-) which had been separated using hydroxyapatite. Gradients 40 rotor
were prepared using 5% and 20% at 38,000 rpm for 17 hr at 20°C.
(w/v)
Assay of Polypyrimidine Segments as Acid-Resistant Acid-Precipitable (ARAP) Radioactivity In previous experiments long pyrimidine acid-hydrolyzed Drosophila DNA were from shorter tracts by ethanol precipitation kaline sucrose gradient sedimentation Straus, and Sederoff, 1975). To facilitate s/s of multiple samples of DNA for their polypyrimidines, a more rapid technique
sucrose
and
tracts in separated and al(Birnboim, the analycontent of has been
0.9 N NaCI-0.1
N NaCH.
Samples
were
centrifuged
in a Beckman
SW
developed. This method is based upon selective precipitation with HCI of long pyrimidine tracts after treatment of DNA with formic acid/diphenylamine. The ARAP method and the standard method (which involves ether extraction and ethanol precipitation) were compared by checking the size of the precipitated polynucleotides on alkaline sucrose gradients (Figure 3a).
Cell 176
The results of such an experiment show that both small and large oligonucleotides were precipitated by the standard method but only large nucleotide fragments were precipitated by the ARAP method (Figure 3a). An additional advantage of the ARAP method is an improved recovery of polypyrimidines, some of which may be lost at the ether extraction step in the standard method (Birnboim et al., 1975). Another demonstration of the ARAP method is shown in Figure 3b. Rapidly reassociating Drosophila DNA was isolated by hydroxyapatite chromatography (Botchan et al., 1971; Kram, Botchan, and Hearst, 1972). Polypyrimidines have a very low Cot, of reassociation when annealed to D. melanogaster DNA (Birnboim et al., 1975); therefore, we expect polypyrimidines to be bound to hydroxyapatite as part of the rapidly reassociated fraction. Both bound and unbound fractions were subjected to ARAP treatment, and the final precipitates were analyzed by alkaline sucrose gradient centrifugation (Figure 3b). As expected, polypyrimidines were found exclusively as part of rapidly reassociating DNA which bound to hydroxyapatite. This experiment also demonstrates the utility of the ARAP procedure as a direct method for quantitating polypyrimidines, since unbound DNA (which lacked polypyrimidines) yielded very little ARAP radioactivity. In another experiment, there was essentially complete recovery of purified 3H polypyrimidines when mixed with unlabeled DNA and carried through the ARAP procedure (data not shown). Because it gave complete recovery of polypyrimidines and little contamination with short pyrimidine tracts, ARAP radioactivity was used as a direct assay for polypyrimidines in subsequent experiments. Distribution of Polypyrimidine-Containing DNA Molecules in CsCl Gradients From the data shown in Figure 3a and other experiments (Birnboim’et al., 1975) we have estimated that the weight average molecular weight of polypyrimidines is 2.48 x 105 daltons. If polypyrimidines were tightly clustered, that is, in adjacent regions in either a cis or tram configuration on duplex DNA molecules, then these DNA molecules could have a buoyant density different from “main-band” DNA and might be separable on a CsCl gradient. If widely distributed throughout the genome, polypyrimidines would be on separate DNA molecules and would likely have little effect on the buoyant density of such molecules provided their size was fairly large (106-l 07 daltons). The buoyant density distribution in CsCl of total DNA (weight average molecular weight in alkaline sucrose, Mwaik = 14 X 106 daltons) and the distribution of polypyrimidine-containing DNA molecules were determined by sedimentation to equilibrium in CsCl (Figure 4). Frac-
B I
20
30 Fraction
Figure 4. Banding midine-Containing
in a CsCl Gradient Molecules
40
6
I
50
number of Total
DNA and of Polypyri-
SH-thymidine-labeled native DNA (&ark = 1.39 x 107 daltons) was centrifuged to equilibrium in a CsCl gradient (see Experimental Procedures), and a portion of each gradient fraction was tested for TCA-precipitable radioactivity (-) and ARAP radioactivity ). The density marker is native E. coli ‘AC(polypyrimidines) (-DNA.
tions were collected and assayed for TCA-precipitable radioactivity (representing total DNA) and ARAP radioactivity (representing polypyrimidine-containing molecules). The major fraction of polypyrimidine-containing molecules are seen to band a position (p g 1.707) which is distinctly heavier than main-band DNA (p = 1.702) and lighter than the E. coli DNA marker (p = 1.710). This experiment indicates that most polypyrimidines are tightly clustered within the genome and behave as a cryptic satellite. Similar experiments were carried out using DNA samples of lower molecular weight (Figure 5). Sheared DNA (MWalk = 1.47 x 106 daltons) and sonicated DNA @Latk = 3.23 x 105 daltons) were centrifuged to equilibrium in CsCl and the positions of total DNA and polypyrimidine-containing molecules were determined. In both cases polypyrimidine-containing molecules banded at a density slightly greater than main-band DNA, although in the case of sonicated DNA the zones were quite broad. These data provide strong evidence that most polypyrimidine segments are clustered-that is, many such segments are grouped together on a single DNA strand.
Polypyrimidines 177
in a Cryptic
Satellite
1 (b)
I.710
(a)
I
I
30
40
50
(TOP--I
20 FRACTION
Figure
5. Banding
in CsCl
-
Gradients
of Native
Sheared
30
DNA and of Polypyrimidine-Containing
Lightly sheared DNA (M,alr = 1.47 x 106 daltons) (a) and sonicated in CsCl gradients as described in Experimental Procedures. Each -o-) and for radioactivity in polypyrimidines (ARAP, -0-). The in alkaline sucrose gradients.
Use of Limited Acid Hydrolysis to Estimate the Extent of Clustering of Polypyrimidines We have developed another approach, different from CsCl banding, to assess the extent of clustering of polypyrimidines in the genome. It is based upon an indirect method for determining the number of purine residues in single-stranded, polypyrimidine-containing molecules. The purine content of a DNA molecule may be estimated from the rate of breakage of the molecule by limited acid treatment, since purine residues are cleaved much more readily than pyrimidine residues, and apurinic sites are broken in alkali. If polypyrimidines were clustered along one strand of DNA, then that strand would have few purines and would be relatively resistant to acid hydrolysis, as compared to “average” DNA molecules which contain about 50% purines. If polypyrimidines were widely interspersed with regions of average composition, then strands to which they were attached would be nearly as sensitive as average DNA molecules. For the experiment, we
40
50
(TOP--)6
NUMBER Molecules
DNA (M Week = 3.23 x 105 daltons) (b) were centrifuged to equilibrium gradient fraction was tested for total radioactivity (TCA-precipitable, inset shows the sedimentation profiles of the two DNA preparations
treated high molecular weight DNA with dilute acid and compared by alkaline sucrose gradient sedimentation the extent to which total DNA and polypyrimidine-containing DNA molecules were degraded (Figure 6). The conditions for limited acid hydrolysis were determined empirically in a preliminary experiment. The profiles of total DNA and polypyrimidine-containing molecules after limited acid treatment clearly indicates that the latter are much more resistant to breakdown. For a more quantitative interpretation, the weight average molecular weights were determined as described in Experimental Procedures. [Although number average molecular weight should be used to estimate strand breaks, we have used weight average molecular weight because number average is strongly biased by small fragments near the top of the gradient. Our conclusion, that there are relatively few purines on polypyrimidine-containing molecules, is not altered by this treatment of the data. Similar treatment of sucrose gradients using weight average molecular weight has been used by Regan, Setlow, and Ley
Cell 178
S 20,
w
8
16
24
32
I
I
I
I
8
6
(1971).] The size of total DNA was initially 13,100 nucleotides long, and this was reduced to 2130 nucleotides after treatment. Polypyrimidine-containing molecules (identified as ARAP radioactivity) were reduced in size from 9830 to 7430 nucleotides. These values indicate that there are only about 8% as many purines in polypyrimidine-containing molecules as there are in “average” DNA (Table 1). Thus it appears that many polypyrimidine segments are grouped together on a single strand of DNA.
Discussion The data presented here and in an earlier report (Birnboim et al., 1975) indicate that an unusual polynucleotide segment can be isolated from Drosophila melanogaster DNA. This segment contains, on the average, 750 pyrimidine nucleotides (34.5% deoxycytidine and 65.5% thymidine) and accounts for about 3% of the thymidine residues of total DNA. By several criteria, including (a) buoyant density in neutral and alkaline CsCl gradients, (b) UV spectral analysis and nucleoside composition, (c) resistance to formic acid/diphenylamine, and (d) ability to anneal at low Cot to homologous DNA, it appears to be an authentic DNA polymer. The large size and unusual composition of these polypyrimidines have provided the basis for an ARAP assay (acid resistant-acid precipitable method), suitable for the analysis of multiple samples from CsCl and alkaline sucrose gradients. The rapid rate of annealing of polypyrimidines with DNA led us to suspect that they had a simple sequence and therefore might be organized as a satellite component. No similar long pyrimidine tracts had been seen previously by Peacock et al. (1973) in the satellites they had examined. We found that polypyrimidine-containing molecules, identified by the ARAP method, were localized principally to a cryptic satellite which bands in CsCl at about 1.707, 5 mg/cmx heavier than main-band DNA. A satellite having a density of 1.705 has recently been identified by Ag+-CsS04 gradient sedimentation and found to direct the synthesis of polypurine complementary RNA (Endow, Polan, and Gall, personal communication). It may be the same as the 1.705 satellite described by Peacock et al.
8
6
Figure 6. Sedimentation in Alkaline Sucrose Gradients DNA and of DNA Treated Briefly with Acid
4
8
Fraction
12
16
20 211
number
of Untreated
The sedimentation profile of total DNA (-0-) and of polypyrimidine-containing molecules (-•-), as assayed by the ARAP method, was determined. In the top figure, untreated DNA was sedimented. In the lower figure, DNA was treated briefly with dilute HCI to introduce a small number of breaks at purine sites before sedimentation. Polypyrimidine-containing molecules changed very little in size, indicating that polypyrimidine segments are tightly clustered (Table 1). Centrifugation was for 16 hr, 24,000 rpm at 20°C.
Polypyrimidines 179
Table
in a Cryptic
1. Calculation
Satellite
of the Number
of Purine
Nucleotides
in Polypyrimidine-Containing Total
Initial Size (nucleotides) Size after
limited
hydrolysis
Breaks
per molecule
a--I ( b Breaks
) per 1000
a
Molecules PolypyrimidineContaining Molecules
DNA
13,100
9830
2,130
7430
5.15
0.32
0.39
0.033
nucleotides
cxE ( Purines
DNA
1
per break 1,282
(“> 2c Purines
per 1000
nucleotides
Cd x e)
500
42
(assumed)
(calculated)
These calculations are based upon the experiment of Figure 6. The sizes of the DNA samples as given in the Table are derived from the weight average molecular weights; although number average molecular weights should theoretically be used, the practical difficulties in using this value have been noted previously (Regan et al., 1971). Isolated polypyrimidines are 750 nucleotides long, as calculated from the weight average molecular weight of their sedimentation profiles.
(1973), although their pyrimidine tract analysis showed no polypyrimidine component. Selective losses of polypyrimidines after formic acid-diphenylamine treatment have been seen previously(Birnboim et al., 1975) and may be responsible for the differences. Several possible models by which polypyrimidines (and their complementary polypurines) may be arranged in the genome are depicted in Figure 7. In model (a), polypyrimidine/polypurine segments are widely scattered throughout the genome, so that not more than one segment is associated with any DNA molecule as normally isolated. In such a case, where the length of the DNA molecule is large relative to the polypyrimidine/polypurine segment, there would be no apparent buoyant density shift. This model is ruled out because a buoyant density shift was observed (Figures 4 and 5). In model (c), the segments are clustered but tandemly inverted in such a way that polypyrimidines and polypurines alternate along each strand. In this model, each strand contains 50% purines and therefore should be as sensitive to partial acid hydrolysis as “average” DNA. This was not the case (Figure 6), making this model unlikely. In model (b), polypyrimidines are completely contiguous-that is, one of the two DNA strands is composed only of pyrimidine residues. If this were the case, polypyrimidines isolated after formic acid-diphenylamine treatment should be as long as the total DNA strand provided nonspecific breaks (at pyrimidine sites) do not occur. A small amount of nonspecific breakage has been observed under these conditions of acid
P
DNA polypyrimidine Polypurine Figure 7. Models in DNA (a) (b) (c) (d)
of the Possible
strand
_
segment
-II
segment
I
Arrangement
of Polypyrimidines
Widely interspersed. Contiguous. Tandemly inverted. Tightly clustered.
Models(b), (c), ments (Figure DNA band as ment (Figure strands which
and(d) are consistent with the CsCl banding experi4 and 5) which show that polypyrimidines in native a “cryptic” satellite. The limited hydrolysis experi6) suggests that polypyrimidines are part of DNA contain relatively few purine residues (model d).
treatment (Birnboim et al., 1975). However, the very mild acid treatment used in the experiment of Figure 6 led to breakage of only about 1 in 1000 purine sites, so no detectable breaks at pyrimidine sites would be anticipated. Figure 6 and Table 1 show
Cl?ll 180
that some change in size of polypyrimidine-containing molecules was detected, indicating that some purines were present. Model(b) is therefore unlikely to be valid. The data are consistent with model (d), in which polypyrimidines are tightly clustered along one strand, separated by, on the average, 33 purines. In an accompanying report (Sederoff, Lowenstein, and Birnboim, 1975), we present evidence concerning the chromosomal location and nucleotide sequence of the polypyrimidine/polypurine segments of D. melanogaster DNA. This component may prove to be particularly useful for evolutionary studies of satellites because a relatively low frequency of purine substitutions in the polypyrimidine regions should be detectable by the methods we have described.
ed by centrifugation and dissolved in 2 ml 0.1 M sodium pyrophosphate at 60°C. For counting, 0.4 ml of 1 N HCI and 10 ml PCS scintillation fluid (Searle Instrumentation) were added. In some experiments, the final pellet was dissolved in 0.2 ml of 0.1 M sodium pyrophosphate and layered on an alkaline sucrose gradient.
Experimental
CsCl Equilibrium Density Gradient Cenfrifugation ,H-thymidine-labeled DNA and polypyrimidine samples, together with an internal E. coli I4C-DNA marker, were centrifuged to equilibrium in CsCl gradients containing either 15 mM Tris-HCI, 3 mM EDTA, 0.5% Sarkosyl (pH 8.0) or 0.1 N NaOH. 10 ml gradients were centrifuged in polyallamer tubes in a Beckman type 65 rotor at 35,000 rpm for at least 60 hr at 20°C, unless otherwise indicated. Fractions were collected dropwise from the bottom, and acidprecipitable counts were determined.
Procedures
Conditions of Cell Culture, Labeling with ‘H-Thymidine, and Preparation of DNA The detailed conditions for growth and labeling of Drosophila melanogaster cells (Schneider’s line 2) and the preparation of SHthymidine-labeled DNA has been described elsewhere (Birnboim et al., 1975). In some experiments, DNA was prepared by the ureahydroxyapatite procedure described by Britten, Graham, and Neufeld (1974). Isolation of Unlabeled Polypyrimidines Ten mg of DNA from cultured ceils was partially hydrolyzed with HCI (0.1 N, lOO”C, 30 min), and the fraction which precipitated on standing at 0°C was treated with 67% formic acid, 2% diphenylamine at 30°C for 40 hr. Polypyrimidines were isolated from the hydrolysate using Sephadex G-50 chromatography as described elsewhere (Birnboim et al., 1975). Base Composition Analysis and UV Spectrum Unlabeled polypyrimidines were dissolved in 0.1 M Tris-HCI (pH 8.0), 1.5 mM MgC12, and hydrolyzed to the nucleoside level using a mixture of venom phosphodiesterase, pancreatic DNAase, and alkaline phosphatase. About 20 nmoles of nucleosides were charged onto a column of Aminex A-6 (0.9 x 25 cm) and eluted with 0.3 M ammonium acetate (pH 5.7) (Singhal, 1972). Nucleosides were monitored with a Chromatronix Model 200 UV detector, and their molar ratio was calculated by reference to standard nucleoside solutions. The UV absorbance spectrum of polypyrimidines was determined using a Beckman Acta V spectrophotometer. Assay for Polypyrimidines as Acid-Resistant Acid-Precipitable (ARAP) Radioactivity 3H-thymidine-labeled DNA and 100 pg carrier DNA were dissolved in 0.25 ml HzO, and 0.75 ml of 90% formic acid-2.7% diphenylamine was added. Hydrolysis was allowed to proceed in the dark at 30°C for 18-20 hr. Long pyrimidine tracts were selectively precipitated from the hydrolysis mixture as follows, To each sample, 0.1 ml of carrier DNA (1 mg/ml), 0.1 ml of 0.1 M sodium pyrophosphate, 0.1 ml of 3 N HCI, and 3 ml of cold ethanol were added in that order. After 30 min at O”C, the precipitate which formed was collected by centrifugation at 30,000 x g for 15 min at 0°C. The pellet was redissolved in 0.8 ml of 0.1 M sodium pyrophosphate, heated at 60°C for 5 min, and reprecipitated at 0°C by the addition of 0.4 ml of 3 N HCI. After 20 min, the sample was centrifuged and the pellet was again redissolved in 0.8 ml of 0.1 M sodium pyrophosphate as before. Following the addition of 0.2 ml of 3 N HCI and a 20 min period of standing at O”C, the final precipitate was collect-
Alkaline Sucrose Gradients for Estimation of DNA Size AH-thymidine-labeled DNA or polypyrimidines were dissolved in 0.2 ml of 0.1 M sodium pyrophosphate and layered over 13.2 ml linear sucrose gradients (5 to 20% w/v in 0.9 N NaCI, 0.1 N NaOH). Gradients were centrifuged in a Beckman SW40 rotor at 20°C for appropriate times and fractionated as described elsewhere (Birnboim, 1972). Sedimentation values were calculated by a procedure similar to that used by McEwen (1967); a program for this purpose using a Hewlett-Packard 9820 A calculator was developed by R. Holford (Atomic Energy of Canada Limited). S values were converted to molecular weights using the relationship of Studier (1265). Weight average molecular weights were calculated as M, = Z qM,/I: ci, where c is radioactivity in cpm and M is calculated molecular weight. [Copies of this report are available from Atomic Energy of Canada Limited, Chalk River: No. AECL-5022.1
Acknowledgments We thank Renate Clynes and J. M. Ostrom for excellent technical assistance. This work was supported in part by an NIH grant. We are grateful to R. Holford of the Health Physics Branch (Atomic Energy of Canada Limited) for developing a computer program for calculating S values and weight average molecular weights. Received
February
6, 1975
References Birnboim,
H. C. (1972).
Biochemistry
Birnboim, H. C., Straus, istry 14, 1643.
17, 4588.
N. A., and Sederoff,
R. R. (1975).
Botchan, M., Kram, R., Schmid, C. W., and Hearst, Proc. Nat. Acad. Sci. USA 68, 1125.
Biochem-
J. E. (1971).
Britten, R. J., Graham, D. E., and Neufeld, B. R. (1974). In Methods in Enzymology, 29, L. Grossman and K. Moldave, eds. (New York: Academic Press), p. 363. Cape, R. (1967). Canada.
Ph.D. Thesis,
McGill
Gall, J. G., Cohen, 33, 319.
E. H., and
Kram, 103.
R., Botchan,
M., and Hearst,
Laird,
C., and
McEwen,
McCarthy,
C. R. (1967).
Polan,
B. (1969). Anal.
University,
Montreal,
M. L. (1971). J. E. (1972). Genetics
Biochem.
Quebec,
Chromosoma J. Mol. Biol. 64,
63, 865.
20, 114.
Peacock, W. J., Brutlag, D., Goldring, E., Appels, R., C. W., and Lindsley, D. L. (1973). Cold Spring Harbor Quant. Biol. 38, 405. Regan, J. D., Setlow, Sci. USA 68, 708.
R. B., and Ley, R. D. (1971).
Riva, S., Barrai, I,, Cavalli-Sforza, Mol. Biol. 45, 367.
L., and
Proc.
Falaschi,
Hinton, Symp.
Nat. Acad. A. (1969).
J.
Polypyrimidines 181
Sederoff, 5, 183.
in a Cryptic
R.,
Lowenstein,
Satellite
L.,
and
Singhal,
R. P. (1972).
Arch.
Studier,
F. W. (1965).
J. Mol. Biol.
Birnboim,
Biochem.
H. C. (1975).
Biophys. II,
Cell
152, 800.
373.
Szybalski, W., and Szybalski, E. H. (1971). In Procedures in Nucleic Acid Research, 2, G. L. Cantoni and D. R. Davies, eds. (New York: Harper and Row), p. 311. Travaglini, 136a.
E., Petrovic,
Travaglini, 419.
E. C., Petrovic,
J., and Schultz, J., and Schultz,
J. (1968). J. (1972).
J. Cell Biol. 39, Genetics
72,