The molecular structure of the transducing particles of Salmonella phage P22. II. Density gradient analysis of DNA

Summary CsCl density gradient analysis showed that the DNA of plaque forming particles ofSalmonella phageP22 is lighter than the host DNA. The DNA of transducing phages exhibits an intermediate density, but close to host DNA. BU labelling of DNA synthesized in the cells after phage infection resulted in a density increase of transducing DNA of about 0.004 gxcm^-3, whereas infectious DNA increased by about 0.045 gxcm^-3. Shearing of isolated DNA molecules from unlabelledP22 lysates demonstrated that transducing DNA consists of two pieces of DNA of different density: 90% stem from the bacterial host whereas 10% are phage DNA and therefore responsible for the BU lable in transducing phages.     

Breakage of Parental DNA Strands in Haemophilus influenzae by 313 nm Radiation after Replication in the Presence of 5-Bromodeoxyuridine

Haemophilus influenzae was labeled with thymidine-³H (dThd), then grown in the presence of 5-bromodeoxyuridine (BrdUrd), and then irradiated with 313 nm light (a wavelength that selectively photolyzes DNA containing 5-bromouracil [BrUra]). Irradiation with 313 nm light induced breaks in the ³H-labeled strands in cells grown with BrdUrd at a much higher frequency than in ¹⁴C-labeled DNA of cells not exposed to BrdUrd. Breakage of the ³H-labeled strands was about 0.6% as efficient as that of fully BrUra-substituted DNA. During growth in the presence of BrdUrd, susceptibility to 313 nm-induced breakage of the ³H-labeled DNA strands increased, reaching a maximum in about one generation, and it decreased to zero during subsequent growth for one generation in medium containing dThd instead of BrdUrd. Heat denaturation of DNA extracted from dThd-³H-labeled cells grown in the presence of BrdUrd eliminated 313 nm-induced breakage of the ³H-labeled strands. It is concluded that breakage of the ³H-labeled DNA strands resulted from reaction with photoproducts in the base-paired, BrUra-containing strands, rather than from photolysis of BrdUrd incorporated into parental strands. It may be possible to utilize the phenomenon of interstrand breakage in physical studies of DNA replication.     

Specific Origin in SV40 DNA Replication

THE oncogenic virus SV40 contains a covalently closed circular DNA molecule of 3 × 10⁶ molecular weight¹. In infected permissive cells, SV40 DNA replicates through a Cairns type intermediate² with the parental strands forming a partially twisted, covalently closed molecule³. We have used a specific bacterial restriction endonuclease⁴ to analyse SV40 DNA replication. The restriction endonuclease of H. influenzae makes double-strand breaks in DNA at specific hexanucleotide sequences^{5, 6} and splits SV40 DNA into eleven fragments, separable by Polyacrylamide gel electrophoresis, ranging in molecular weight from 6.5 × 10⁵ (about 20% of the molecule) to 7.4 × 10⁴ (about 2.5% of the molecule). The largest eight fragments are present in the digest in amounts equimolar with the starting DNA⁴. Therefore, by digesting labelled replicating SV40 DNA and newly completed DNA and measuring the relative yield of each fragment, we could determine whether a particular region of the DNA is synthesized first or last and also estimate the time needed to replicate one molecule completely.     

In vitro incorporation of tritium into native DNA

An exchange method is described for producing tritium-labeled native DNA in vitro with minimal physical damage to the DNA. Tritium-labeled calf thymus DNA prepared in this way has a specific activity of about 100 μCi/mmole of nucleotide (i.e., about 2 × 10⁸ dpm/mmole). Sedimentation velocity at neutral and alkaline pH indicate that the product has an average of two single strand breaks per duplex molecule of molecular weight 6 × 10⁶ daltons. The optical and thermal denaturation properties of the product are those of native DNA. The method should be particularly useful for labeling DNA from organisms that cannot be labeled conveniently in vivo.     

Ribosomal RNA cistrons in Euglena gracilis

Euglena gracilis chloroplasts contain about 12 fg DNA of average density 1.686 g cm^?3 and 1.7 pg RNA. The large (1.1 × 10⁶ mol. wt) and small (0.56 × 10⁶ mol. wt) ribosomal RNA components are coded for by separate cistrons, both of which band at a density of 1.696 g cm^?3 in a CsCl gradient. About 6% of the chloroplast DNA codes for rRNA indicating that there are 240 cistrons for rRNA in each chloroplast or about three to six cistrons per chloroplast genome. Similar studies with rRNA from cytoplasmic ribosomes indicate that the cistrons for cytoplasmic rRNA band at a density of 1.716 g cm^?3, denser than that of the main-band DNA, and that there are 1000 cistrons for cytoplasmic rRNA per cell. Fractionation of E. gracilis DNA on CsCl gradients and subsequent hybridization experiments, as well as melting curves of DNA-RNA hybrids, show that chloroplast rRNA does not anneal specifically with either the cistrons for cytoplasmic rRNA or any DNA in the dark-grown cell, in contrast to those results found in some higher plants.     

Nucleases in chloroplast of barley

Two barley chloroplast nuclease fractions were separated by the affinity chromatography and gel electrophoresis. Both were about 2 times more active to RNA than to native DNA and about half as active to denaturated DNA as to native DNA. Both fractions were as active to UV-irradiated (270 J m^-2) native DNA as to intact DNA but their action was inhibited by apurinic sites. The enzyme activities were inhibited by high concentrations of EDTA, NaCl, Mn²⁺, Ca²⁺, Zn²⁺ ions and by N-ethylmaleimide. They do not require Mg²⁺ ions but are stimulated or at higher concentration inhibited by their presence. Both RNase and DNase were active over a wide pH range (5.5–9), the optimum for DNase action in the presence of Mg²⁺ being 6.5, for RNA decomposing activity at pH 8.0. As no mononucleotides were detected in acid soluble form, it seems likely that DNase acts in the endonucleolytic way.     

DNA of ciliated protozoa: DNA sequence diminution during macronuclear development of oxytricha

We have measured the reassociation kinetics of DNA from the micronucleus and from the macronucleus of the hypotrichous cillate Oxytricha. The micronuclear DNA reassociates with at least a two-component reaction, indicating the presence of both repeated and non-repeated sequences. The kinetic complexity of micronuclear non-repeated DNA is in the range of 2 to 15 × 10¹¹ daltons; the haploid DNA content of the micronucleus is 4 × 10¹¹ daltons (0.66 pg), measured microspectrophotometrically. The DNA of the macronucleus reassociates as a single second-order reaction, with a kinetic complexity of 3.6 × 10¹⁰ daltons. A comparison of the kinetic complexities of micronuclear and macronuclear DNAs suggest a 5 to 30 fold reduction in DNA sequence complexity during the formation of a macronucleus from a micronucleus. Macronuclear DNA is in pleces with an average molecular weight of 2.1 × 10⁶ daltons. Since the kinetic complexity of macronuclear DNA is 3.6 × 10¹⁰ daltons, the macronucleus must contain about 17,000 different kinds of DNA pieces.Each macronucleus contains 3.5 × 10¹³ daltons (58 pg) of DNA, indicating that each sequence must be present about 1000 times per macronucleus or 2000 times per cell.     

Spacing of polypyrimidine regions in mouse DNA as determined by poly(adenylate,guanylate) binding

The DNA from many higher organisms contains unexpectedly long runs of pyrimidine nucleotides (polypyrimidines). In mouse cells, these range in size from about 25 to 250 nucleotides and account for about 0.5% of the total DNA. They contain both thymine and cytosine in close to equimolar amounts but few or none are pure thymine or cytosine runs. When native DNA from mouse cells was banded in a CsCl gradient, polypyrimidines were part of DNA molecules of all base compositions. This indicates that, unlike Drosophila melanogaster DNA (Birnboim & Sederoff, 1975), they are not clustered as a satellite component. A procedure for isolating fragments of single-stranded DNA which contain polypyrimidines has been developed. Polypyrimidine sites in DNA are complexed with poly(A,G) and the DNA/poly(A,G) complex is bound to hydroxyapatite. The percentage of DNA which binds is a function of its chain length. A maximum of 60% of the DNA was bound, at lengths of > 15 × 10³ bases². We conclude that polypyrimidine sites are distributed throughout the majority of the mouse genome at intervals of 12 to 15 × 10³ bases.     

Polypyrimidine segments in drosophila melanogaster DNA: I. Detection of a cryptic satellite containing polypyrimidine/polypurine DNA

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 CsCI 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 10⁵-10⁷ 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 CsCI gradients was consistent with a single-stranded DNA polymer of this composition.     

Nascent DNA in nucleosome like structures from chromatin

We have used chromatin sensitivity to cleavage by micrococcal nuclease as a probe for differences between chromatin containing nascent DNA and that containing bulk DNA. Micrococcal nuclease digested the nascent DNA in chromatin of swimming blastulae of sea urchins more rapidly to acid-soluble nucleotides than the DNA of bulk chromatin. A part of the nascent DNA occurred in micrococcal nuclease-resistant structures which were either different from, or temporary modifications of, the bulk nucleosomes. This was inferred from the size differences between bulk and nascent DNA fragments in 10% polyacrylamide gels after micrococcal nuclease digestion of nuclei from a mixture of ¹⁴C-thymidine long- and ³H-thymidine pulse-labeled embryos. Bulk monomer and dimer DNA fragments contained about 170 and 410 base pairs (bp), respectively, when 18% of the bulk DNA had been rendered acid-soluble. At this level of digestion, “nascent monomer DNA” fragments of about 150 bp as well as 305 bp “large nascent DNA fragments” were observed. Increasing levels of digestion indicated that the large nascent DNA fragment was derived from a chromatin structure which was more resistant to micrococcal nuclease cleavage than bulk dimer chromatin subunits. Peaks of ³H-thymidine-labeled DNA fragments from embryos which had been pulse-labeled and then chased or labeled for several minutes overlapped those of ¹⁴C-thymidine long-labeled monomer, dimer and trimer fragments. This indicated that the chromatin organization at or near the replication fork which had temporarily changed during replication had returned to the organization of its nonreplicating state.     

The major cause of inactivation and mutation by 4-Nitroquinoline 1-Oxide in Escherichia coli: Excisable 4NQO-purine adducts

The potent carcinogen, 4-nitroquinoline 1-oxide, is known to mimic the biological effects of ultraviolet light on various living organisms. We conclude that the 4NQO² effects on Escherichia coli are mostly due to covalent binding of 4NQO to DNA producing 4NQO-guanine and 4NQO-adenine adducts in a ratio of about 4:1 without repair and about 7:1 after repair. This is based on the following experimental results. From E. coli cells treated with [³H]4NQO, DNA was extracted and subjected to radiochromatography. We detected two peaks of 4NQO-guanine adduct, one peak of 4NQO-adenine adduct and a peak due to 4-aminoquinoline 1-oxide released from a labile fraction of 4NQO-guanine adducts during acid hydrolysis of DNA before streaking it on paper for chromatography. These four kinds of 4NQO-purine adducts disappeared from DNA of the normal strain at almost the same rate, about 85% in 60 minutes by post-incubating in nutrient broth, but these adducts did not disappear for the uvrA⁻ derivative lacking the excision-repair ability for ultraviolet-induced pyrimidine dimers, except for slight disappearance of 4AQO-releasing adduct. The number of DNA lesions per genome of the uvrA⁻ strain at 37% survival was found to be nearly equal between the 4NQO-purine adducts (~200 lesions) and pyrimidine dimers (~100 lesions). These findings at the molecular level quantitatively parallel the previous findings at the cellular level that the uvrA⁻ strain is about 25 to 30 times as sensitive as its parental strain to killing and mutation by either 4NQO or ultraviolet light. The unique characteristics of 4NQO-purine products are discussed in relation to the mutational specificity of 4NQO and the more-than-supposed complexity of excision repair for DNA.     

Increase in DNA replication sites in cells held at the beginning of S phase

CHO cells were pulse labeled with ³H-thymidine after synchronization and blockage at the beginning of S phase for various intervals. The distribution of initiation sites for DNA replication and rates of chain growth were measured in autoradiographs prepared from these cells. Origins used for replication are widely distributed at or near the beginning of S phase, but usable origins increase continuously for many hours when FdU is used to block the synthesis of thymidylate. Potential origins are located about four microns apart, but in normal replication in these fibroblasts only one in 15 to 20 potential origins are used for initiation. On the other hand, when cells are held at the beginning of S phase for 12–14 h, about one-half of the potential origins are activated in part of the DNA and utilized when the cell is released from the block by supplying ³H-thymidine (10^–6M). Chain growth during a short pulse decreases with time of the blockage at what appears to be a linear rate. However, cells can replicate long continuous stretches of their DNA with only 2×10^–8M thymidine available in the medium for several hours when synthesis is blocked by FdU. The total amount of DNA replicated is, however, much less than when a concentration of 10^–6 M thymidine is supplied for the same period. The origins which are finally used under any experimental condition appear to be a random sample of the total potential origins which are distributed in a regular repeating sequence along the DNA at about 12 kilobase intervals.     

DNA-induced endocytosis upon local microinjection to giant unilamellar cationic vesicles

We suggest a novel approach for direct optical microscopy observation of DNA interaction with the bilayers of giant cationic liposomes. Giant unilamellar vesicles, about 100 μm in diameter, made of phosphatidylcholines and up to 33 mol% of the natural bioactive cationic amphiphile sphingosine, were obtained by electroformation. “Short” DNAs (oligonucleotide 21b and calf thymus 250 bp) were locally injected by micropipette to a part of the giant unilamellar vesicle (GUV) membrane. DNAs were injected native, as well as marked with a fluorescent dye. The resulting membrane topology transformations were monitored in phase contrast, while DNA distribution was followed in fluorescence. We observed DNA-induced endocytosis due to the DNA/lipid membrane local interactions and complex formation. A characteristic minimum concentration (C _endo) of d-erythro-sphingosine (Sph⁺) in the GUV membrane was necessary for the endocytic phenomenon to occur. Below C _endo, only lateral adhesions between neighboring vesicles were observed upon DNA local addition. C _endo depends on the type of zwitterionic (phosphocholine) lipid used, being about 10 mol% for DPhPC/Sph⁺ GUVs and about 20 mol% for SOPC/Sph⁺ or eggPC/Sph⁺ GUVs. The characteristic sizes and shapes of the resulting endosomes depend on the kind of DNA, and initial GUV membrane tension. When the fluorescent DNA marker dye was injected after the DNA/lipid local interaction and complex formation, no fluorescence was detected. This observation could be explained if one assumes that the DNA is protected by lipids in the DNA/lipid complex, thereby inaccessible for the dye molecules. We suggest a possible mechanism for DNA/lipid membrane interaction involving DNA encapsulation within an inverted micelle included in the lipid membrane. Our model observations could help in understanding events associated with the interaction of DNA with biological membranes, as well as cationic liposomes/DNA complex formation in gene transfer processes. Received: 18 April 1998 / Revised version: 6 August 1998 / Accepted: 7 August 1998     

Characterization of cytoplasmic and nuclear genomes in the colorless alga Polytoma

DNA isolated from purified nuclei of Polytoma obtusum has a buoyant density of 1.711 g/ml in CsCl, a Tm of 91.3° C in SSC, and a G + C content of 52.5% as determined by base composition analysis. Thermal dissociation and reassociation studies indicated that this nuclear DNA contains a considerable amount of heterogeneity. Under appropriate reannealing conditions for denatured DNA, about 15% of the DNA reannealed to form a satellite peak at a density of 1.711 g/ml within one hour. Native DNA fractions of different average buoyant densities, ranging from 1.723 to 1.708 g/ml were also obtained in a preparative CsCl gradient, indicating the presence of intermolecular heterogeneity at a molecular size of 8.5×10⁶ daltons. The nuclear DNA reassociated as three distinct classes. The very fast species constituted about 20 % of the total hyperchromicity, the class of intermediate rate comprised roughly 10% of the nuclear DNA, while the remaining 70% consisted of unique sequences. The haploid genome set was estimated by renaturation kinetics studies to contain 5.0×10¹⁰ daltons of DNA or 7.5×10⁷ nucleotide pairs. The analytical complexity of the total nuclear genome was found to be 9.35×10¹⁰ daltons, thus indicating that vegetative cells of P. obtusum are diploid.     

Aflatoxin B1 and DNA Adducts. Proposed Model for Surface Noncovalent and Covalent Complexes with N(7) of Guanine. II.

Abstract

An alternate model for surface noncovalent and surface covalent binding of aflatoxin B₁ to N(7) of guanine in DNA is proposed. This model considers the out-of-plane motions of C(8) of aflatoxin B₁ in those interactions.

The covalent intercalated fit of aflatoxin B₁ into DNA arises from steric adjustments made by DNA at the covalent intercalation site as well as local strain in the bond angles about N(7) of guanine and C(8) of aflatoxin B₁. The bond angle about N(7) deviates modestly from the sp² value toward the sp³ value.

This study suggests that the surface covalent aflatoxin B₁ -DNA complex serves only a minor role in aflatoxin's precarcinogenic interaction with DNA and is a likely correctable error.     

Linear dichroism studies of nucleic acids. II. Calculation of reduced dichroism curves of A-and B-form DNA

We have calculated the uv linear dichroism for the A- and B-forms of DNA using π-π* transition moments and band components determined from the free DNA bases. The reduced dichroism (LD^R) as a function of wavelength is estimated is in the 220–300-nm region, for both the oriented-gas model and a simple exciton model. For B-form DNA, LD^R is obtained to ?1.48S (S being the orientation factor) over the whole wavelenth region by both models. For A-form DNA, LD^R is not constant, but changes monotonically from about ?1.15S at 220 nm to about ?1.35S to ?1.45S at 300 nm, depending on base combination and degree of interaction (?1.35S for the oriented gas). It is emphasized that a common assumption of a single “effective” transition moment of the principal band at 260 nm may not generally be made because of the extensive overlap of differently polarized bands. The possibility of using the reduced dichroism curve for characterizing the secondary structure of DNA is discussed.     

Extent and location of DNA synthesis associated with a class of Rec-mediated recombinants of bacteriophage lambda

The extent and location of DNA synthesis associated with Rec recombination of a lambda phage mutant has been determined approximately for recombinants arising under conditions that restrict DNA duplication. The mutant bio1 contains a substitution in its DNA, and nearly all phage maturing under these conditions have undergone a recombination event within a short region in or near the inserted DNA. Density labeled phage bio1 were used to prepare a lysate under these conditions and the extent of new DNA synthesis was determined by analyzing the density of the progeny phage. On the average, about 6% of the phage chromosome was resynthesized in such a cross.DNA was extracted from bio1 phage crossed under similar conditions in the presence of ³²PO₄. The position of incorporated ³²PO₄ was determined by cleaving the DNA with EcoRI restriction endonuclease and resolving the resulting fragments by electrophoresis on agarose gels. The fragment found to have the most newly synthesized DNA and the highest average amount of synthesis per nucleotide contains the bio1 insertion near its left end and the “hot spot” for Rec-mediated recombination near its center. It appears that in these crosses recombination-associated DNA synthesis is localized about the region of the Rec-mediated recombination event.     

Physical properties of cytoplasmic membrane-associated DNA

Some of the physical properties of a cytoplasmic membrane-associated DNA isolated from a diploid human lymphocyte cell line have been examined. Cytoplasmic membrane-associated DNA extracted from lymphocytes labeled with either [³H]or [¹⁴C]thymidine had a specific activity lower than nuclear DNA extracted from the same cells. Analysis of cytoplasmic membrane-associated DNA in the electron microscope shows that the molecules are linear and have a mean length of 1·75 μm; the average sedimentation coefficient of this DNA is 16·6 S, which corresponds to a molecular weight of 4·2×10⁶. Cytoplasmic membrane-associated and nuclear DNA band at identical positions in both neutral and alkaline CsCl gradients with buoyant densities of 1·699 g/ml and 1·752 g/ml, respectively. Native cytoplasmic membrane-associated DNA is double-stranded and has a mole fraction of guanine plus cytosine of 40± l %. Sheared, denatured cytoplasmic membrane-associated DNA reassociates as two distinct fractions whose rates of reassociation differ by about four decades: the complexity of the reassociation of this DNA tends to rule out the possibility that it arises from either mycoplasmal or viral contamination of our cell cultures. The slowly reassociating fraction of cytoplasmic membrane-associated DNA reassociates about ten times faster than the unique sequences of nuclear DNA. This could represent potential genetic information for about 100,000 diverse genes of 1000 nucleotide pairs each. At present the function of cytoplasmic membrane-associated DNA in these cells is unknown.     

High-affinity microtubule protein-higher organism DNA complexes. Many-fold enrichment in repetitive mouse DNA sequences comprised of satellite DNAs

We have examined aspects of the interaction of cycled microtubule protein preparations with ³⁵S-labeled mouse DNA tracer in a competition system with unlabelled competitor E. coli or mouse DNA. The nitrocellulose filter binding assay was used to measure interaction by scintillation counting. DNA molecular weight affected the levels of filter retained ³⁵S-labelled mouse tracer DNA. Filter retention levels increased if ³⁵S-labelled mouse DNA tracer size was increased, and the filter binding level decreased if competitor DNA size was increased. There was a sizeable, reproducible difference in the ³⁵S-labelled mouse DNA tracer binding level of about 1% when E. coli or mouse DNA competitors were compared. Mouse DNA more effectively competed with ³⁵S-labelled mouse DNA for microtubule protein binding than did E. coli DNA, suggesting that a small class of higher-organism DNA sequences interacts very strongly with microtubule protein. From other studies we know this to be the MAP fraction (Marx, K.A. and Denial, T. (1984) in The Molecular Basis of Cancer (Rein, R., ed.), Alan R. Liss, New York, in the press; and Villasante, E., Corces, V.G., Manso-Martinez, R. and Avila, J. (1981) Nucleic Acids Res. 9, 895–908). We find that this difference in competitor DNA strength is qualitatively similar under high-stringency conditions (0.5 M NaCl, high competitor [DNA]) we developed for examining high-affinity complexes. Under high-stringency conditions we isolated 1.2% and 0.6% of ³⁵S-labelled mouse DNA at 4200 and 350 bp respective sizes as nitrocellulose filter bound DNA-protein complexes. At both molecular weights these high-affinity DNA sequences, isolated from the filters, were shown to be significantly enriched in repetitive DNA sequences by S₁ nuclease solution reassociation kinetics. The kinetics are consistent with about a 4-fold mouse satellite DNA enrichment as well as enrichment in other repetitious DNA sequence classes. The high molecular weight filter-bound DNA samples were sedimented to equilibrium in CsCl buoyant density gradients and found to contain primarily mouse satellite DNA density sequences (1.691 g/cm³) with some minor fractions at other density positions (1.670, 1.682, 1.705, 1.740, 1.760 g/cm³) similar to those observed by our laboratory in previous investigations of micrococcal nuclease-resistant chromatin (Marx, K.A. (1977) Biochem. Biophys. Res. Commun. 78, 777–784). That the high-affinity microtubule-bound DNA was some 3–5-fold enriched in mouse satellite sequences was demonstrated by its characteristic BstNI restriction enzyme cleavage pattern     

The ribosomal cistrons of the water mold Achlya bisexualis

Total DNA was extracted from vegatative mycelia of the water mold Achlya bisexualis. Fractionation of the DNA in CsCl gradients resulted in two components: a major component with a buoyant density of 1.697 g cm^?3 and a minor component with a density of 1.685 g cm^?3. The minor component has been identified as mitochondrial DNA based on extractions from isolated mitochondria and Triton X-100 washed nuclei. Detergent washing of the nuclei yielded DNA in which the mitochondrial DNA component was absent, while the isolated mitochondrial preparations contained DNA enriched in the 1.685 g cm^?3 component. Hybridization studies of A. bisexualis DNA to rRNA show that the ribosomal cistrons have a buoyant density coincident with that obtained with the nuclear DNA. In addition, preliminary evidence indicates that the mitochondrial DNA does not hybridize to the cytoplasmic RNA under the conditions used for this study. Ribosomal RNA hybridized to about 0.65% of the total DNA.