Synthesis of a Deoxyribonucleic Acid Sequence Complementary to Ovalbumin Messenger Ribonucleic Acid and Quantification of Ovalbumin Genes*

SUMMARY DNA was synthesized in vitro from a template of purified ovalbumin messenger RNA. The DNA product is complementary to a homogeneous sequence specific for ovalbumin. The amount of ovalbumin sequence in hen oviduct and hen liver DNA was measured by hybridization experiments and found to be 2 copies per diploid genome in both tissues, indicating that there is no selective amplification of ovalbumin genes in the tissue specialized for obalbumin synthesis. nucleic acid quantity of information

Both DNA (l-3) and RNA (4) polymerases can be used to synthesize nucleic acid sequences complementary to hemoglobin messenger RNA.
Such sequences, specific for a given polypeptide, can be used in molecular hybridization experiments to estimate the quantity of information coding for that polypeptide in a given population of nucleic acid molecules (5-7). We have previously reported the purification of messenger RNA for chicken ovalbumin by selective immunoadsorption of ovalbumin polyribosomes and subsequent adsorption of messenger RNA to nitrocellulose filters (8). The present communication describes the preparation of DNA complementary to this messenger RNA using the RNA-directed DNA polymerase of Rous sarcoma virus (9). Analysis of this DNA by molecular hybridization experiments indicates that the enzymatic product represents a homogeneous set of nucleotide sequences specific for * This work was supported in part by Research Grant GM 14931 ovalbumin, and that 2 copies of these sequences per diploid cell are present in both hen oviduct and hen liver.
We conclude that the extensive synthesis of ovalbumin by oviduct cells cannot be attributed to gene amplification.

MATERIALS AND METHODS
Most of the materials used have been described elsewhere (8) Preparation of Purijied Ovalbumin mRNA-Details of this procedure and controls on the homogeneity of the product have been described at length elsewhere (8). Further documentation of its homogeneity is provided below (see "Results"). Polysomes were isolated from adult hen oviduct by centrifugation to a sucrose cushion (10). These polysomes (10 As60 units per ml) in polysome buffer (0.025 M NaCl-0.005 M ?\IgC&-0.025 RI Tris-HCl, pH 7.6, containing 100 pg of heparin per ml) were incubated with purified goat antiovalbumin antibody (0.5 mg per ml) for 45 min at 4". A matrix of cross-linked ovalbumin (ZOO mg of matrix per ml of reaction mixture) was then added to the reaction mixture, and the mixture was incubated for 45 min more at 4" with constant stirring.
The mixture was then centrifuged at 4" for 10 min at 6,000 rpm in the Sorvall HB4 rotor, the supernatant fluid was discarded, and the matrix was washed three times with 0.5 M sucrose, 0.15 M NaCI, 1% Triton X-100, and 1% sodium deoxycholate in polysome buffer (1 ml of buffer per 100 mg, wet weight, of matrix), then one time with polysome buffer alone.
To elute the immunoadsorbed polysomes, a buffer containing 0.01 M Tris-HCI, 0.05 M EDTA, pH 7.5, and 100 pg of heparin per ml was added (0.5 ml/100 mg, wet weight, of matrix) and incubated for 15 min at 4' with constant stirring. The preparation was centrifuged as before and the supernatant fluid was saved.
Two more elutions were performed and the three supernatant fluids were pooled. The pooled supernatant fractions were made 0.1 M in NaCl and precipitated with 2 volumes of et.hanol at -20" for 12 hours.
The precipitate was then dissolved in 0.5% SDS'-0.02 M sodium acetate-O.005 M EDTA, pH 5.0, at a final concentration of 10 A260 per ml. One-milliliter samples were layered on 11.5.ml 5 to 20% sucrose gradients in the same buffer and centrifuged at 40,000 rpm for 6 hours at 20" in a Spinco SW 41 rotor.
The gradient was pumped through a flow cell in a Gilford recording spectrophotometer, and all material sedimenting more rapidly than the tRNA and SDS-treated proteins was collected and precipitated with ethanol three times as described above.
This RNA was then adsorbed to Millipore filters according to a modification (8) of the procedure of Brawerman et a?. (11) to selectively enrich for mRNA.
We have previously documented the efficacy of our procedure for preparing ovalbumin-specific polyribosomes (8,10,12,13 SDS was added to 0.5% and the material was homogenized with three more strokes. Pronase (selfdigested in 0.02 M Tris-0.01 M EDTA, pH 7, for 2 hours at 37") was added to a final concentration of 200 pg per ml and the preparation was incubated overnight at 37". More SDS was added to a final concentration of 1 yO, the material was extracted twice with an equal volume of phenol (previously saturated with 0.02 M sodium acetate), and 2 volumes of ethanol were added. After 48 hours at -20" the material was centrifuged for 20 min at 10,000 rpm in a Sorvall HB4 rotor, and the precipitate was dissolved in 0.001 M Tris-HCI-0.01 M EDTA, pH 7.0. Ribonuclease (Worthington, DNase-free, previously boiled for 10 min at 10 mg per ml) was added to a final concentration of 100 pg per ml and the preparation was incubated overnight at 37". The material was extracted three times with phenol as described above, and the aqueous phase was dialyzed exhaustively against 0.015 M NaCl-0.  in duplicate with and without S1 nuclease at 50" for 2 hours.
The amount of nuclease activity added was always in IO-fold excess over the amount required to obtain complete hydrolysis of the susceptible substrate. At the end of the incubation period 40 pg of native calf thymus DNA were added as carrier, the samples were precipitated with trichloroacetic acid and filtered, and the amount of acid-precipitable radioactivity was measured.
For further details of this procedure, see Leong et al. (16).
The quantity of nuclease-resistant material is expressed as a percentage of the acid-precipitable radioactivity in the samples incubated without S1 nuclease. This mixture was incubated for 16 to 18 hours at 37", then SDS was added to 0.5% and the mixture was incubated at 37" for 5 min more.
HeLa cell RNA, 20 pg prepared as described previously (9), was added as carrier along with sodium acetate to 0.2 M and ethanol to 67 '% (v/v).
The sample was allowed to precipitate at -20" for at least 6 hours.
After centrifugation at 14,000 X g for 20 min, the precipitate was dissolved in 200 ~1 of lop4 M EDTA, boiled RNase was added to 100 pg per ml, and the sample was incubated for 1 hour at 37" to destroy the template RNA and disrupt DNA-RNA hybrids (16). At the end of this time the sample was diluted into 2 ml of 0.01 M sodium phosphate, pH 6.8, and fractionated on hydroxyapatite as described above. The single stranded material, about 70y0 of the total DNA synthesized, was then passed through a column (40 X 0.9 cm) of Sephadex G-50 (coarse) in 0.3 M NaCl, 0.01 M Tris-HCI, pH 8.1, 0.001 M EDTA, and 0.1 y0 SDS.
The void volume was collected, ethanol was added to 67% (v/v), and the sample was precipitated at -20".
Double stranded DNA was synthesized under identical conditions except that actinomycin D was omitted from the enzymatic reaction.
After elution from hydroxyapatite, the fraction eluting in 0.4 M sodium phosphate was passed through the Sephadex column.
This double stranded material was found to contain a fraction (about 40% of the total) resistant to denaturation. This resistant material was removed by boiling the double stranded DNA for 10 min in 0.003 M EDTA, pH 7, adsorbing the DNA to hydroxyapatite again, and eluting the single stranded material as described above.
The amount of DNA synthesized in vitro was never sufficient for optical measurement of the mass of DNA present. From the specific activities of the nucleoside triphosphates used in the synthesis of this DNA, and from our estimates of the base ratios of the DNA product (Table II, below), we estimate the specific activity of the DNA product to be about 8000 cpm per ng. Assay of Ovalbumin mRNA Activity in Cell-free Protein-synthesizing System-This technique, consisting of immunoprecipitation of ovalbumin synthesized in a rabbit reticulocyte lysate system, was as described previously (18) except that isoleucine was used as the labeled amino acid (8).
Isolation of Hen Oviduct Monosomal RNA--Hen oviduct monosomes were purified by rate zonal centrifugation in a 0.5 to 1.5 M sucrose gradient (19) and deproteinized by centrifugation in an SDS-sucrose gradient as described previously (8). This RNA was assayed in the reticulocyte lysate protein synthesis system (see above) and found to be 3% as active at ovalbumin synthesis per unit mass of RNA as hen oviduct polysomal RNA, and about 0.1 y0 as active as our purified messenger RNA (data not shown).
DNA-DNA Hybridization-DNA samples were denatured by boiling for 10 min in 0.003 M EDTA, pH 7. Salt was then added to a final concentration of 0.6 M NaCl, 0.02 M Tris-HCI, pH 7.0, and 0.002 M EDTA.
Samples were overlayered with mineral oil to prevent evaporation, incubated at 68" for various times, then assayed for secondary structure by S1 nuclease (see above). The percentage hybridized was plotted against the C&Z In these calculations we assumed an average value of 309 g of DNA nucleotides per mole.

RNA-DNA
Hybridization-To denature the DNA and destroy any ribonuclease present,3 single stranded DNA product was boiled for 10 min in 0.3 M NaOH-0.003 M EDTA, then neutralized with HCl.
RNA was added, then buffer to achieve the nucleic acid concentration desired and a final salt concentrat.ion of 0.3 M NaCl, 0.01 M Tris-HCl, pH 7.0, and 0.002 M EDTA. All hybridization reactions (including controls on self-annealing of single stranded DNA product) were performed in the presence of 1 mg of HeLa cell RNA per ml, added to reduce the effect of any trace of ribonuclease on the very small amounts of RNA present.
The DNA did not react with HeLa RNA in appropriate pilot experiments (Table IV, below). Samples were overlayered with mineral oil, incubated at 68" for up to 24 hours, and analyzed for secondary structure by the X1 nuclease assay (see above).
The reciprocal of the percentage double strand was plotted versus the reciprocal of the C,t.
In these calculations we assumed an average molecular weight of 321 g of RNA nucleotides per mole.
Since all RNA-DNA hybridization reactions were done under conditions of great RNA excess (20-to lO,OOOfold, as indicated), the reactions should be pseudo-first order (21) and therefore a hybridization reaction of a homogeneous population of molecules should plot as a straight line. The intercept of this line with the ordinate (at l/C& = 0 or CJ = m ) represents 2 Cot is defined as the product of concentration of DNA (in moles of nucleotides per liter) and time (in seconds), multiplied by a constant factor of 5 to correct to standard Na+ concentration (0.12 M sodium phosphate) in the annealing mixture (20). C,t is the product of concentration of RNA (in moles of nucleotides per liter) and time (in seconds) multiplied by 2.3 to correct to standard Na+ concentration (0.12 M sodium phosphate) in the annealing mixture (20). Cot 1/2 or C,f, 12 is the Cot or C,t (in mole .s per liter) at which the hybridization of a given population of nucleic acid molecules is one-half completed.
the reciprocal of the percentage of the probe that is double stranded when that population of molecules has reacted for an infinite time (21).
Since RNA was present in great excess the concentrations of the RNA species determined the kinetics of the reaction, and the homogeneity and complexity measured by these reactions are those of the RNA, not the DNA species. This type of analysis of RNA-DNA hybridization reactions is discussed more fully by Birnstiel et al. (21). Preparation of Labeled, Unique Sequence Chicken DNA-Chick embryo fibroblasts were cultured for 48 hours in the presence of 2.5 &i of [l%]thymidine per ml. DNA was then prepared by adsorption and elution from hydroxyapatite essentially as described by Britten et al. (22), except that the lysed cells were sheared by passage through a 26.gauge needle. The double stranded DNA was then sheared at 50,000 p.s.i. as described above and dialyzed against 0.003 M EDTA.
This material was precipitated by the addition of ethanol, then denatured and reannealed as described above to a Cot value of 6 X lo2 mo1e.s per liter.
The material was fractionated by hydroxyapatite (see above), and the single stranded material was dialyzed against 0.003 M EDTA and precipitated by addition of ethanol. This material was considered to be unique sequence chicken DNA; it had a specific activity of 10,000 cpm per pg.

RESULTS
Template Specijicity of RSV DNA Polymerase and Characterization of DNA Product-Purified ovalbumin mRNA supports DNA synthesis at a rate even greater than that obtained with RSV 70 S RNA (Table I).
In a 50.~1 reaction mixture containing 65 ng of purified ovalbumin mRNA, 80,000 cpm or about 10 ng of DNA were synthesized.
Preincubating the purified ovalbumin mRNA with RNase destroyed the template activity, indicating that RNA, not a DXA contaminant, was directing the synthesis.
The requirement for oligo(dT) is in agreement with the findings of others for globin mRNA (l-3) and presents further evidence for the existence of an adenine-rich sequence in ovalbumin mRNA (8, 23). The enzyme had a low but detectable level of endogenous DNA synthesis.
This synthesis was presumably the result of a small amount of viral RNA in the enzyme preparation, since added ribonuclease reduced the endogenous activity several-fold. Since this background level of DNA synthesis was only 1 to 2% of that in the reaction with added RNA, no steps were taken to further purify the enzyme, although such techniques have been developed (9). We have previously presented evidence (8) that our purified ovalbumin mRNA still contains rRNA. Although other studies (1, 9) have shown that rRNA is not transcribed by DNA polymerase from tumor viruses, we tested whether hen rRNA might serve as template under our conditions of synthesis. RNA from hen oviduct monosomes was isolated by sucrose gradient centrifugation (see "Materials and Methods") and tested for relative ability to direct DXA synthesis in vitro (Table I).
This monoso-ma1 RNA had a higher relative concentration of rRNA than the purified ovalbumin mRNA but much less messenger RNA (see "Materials and Methods" and Ref. 8). The amount of DNA synthesis (including endogenous background synthesis) was only 3% of the amount of DNA synthesis in the presence of purified ovalbumin mRNA.
A reaction mixture containing both purified ovalbumin mRNA and monosomal RNA produced approximately the same amount of synthesis as purified mRNA alone, indicating that the low synthesis in the presence of monosomal RNA was not the result of an inhibitor in this fraction. From these experiments we conclude that less than 3% of the DNA synthesis in the presence of purified ovalbumin mRNA is directed by ribosomal or viral RNAs. The single stranded enzymatic product has a range of sedimentation coefficient centering about 7 S (Fig. 1). This indicates an average chain length of about 200 nucleotides, and compares favorably with the size of DNA transcribed from larger RNA templates under similar conditions (see "Discussion"). The nucleotide composition of the enzymatic product was determined with DNA synthesized in the presence of all four radioactive nucleoside triphosphates (Table II). The base ratios of ovalbumin mRNA are not known; therefore we cannot state whether the nucleotide composition of the DNA product is appropriate for a sequence complementary to this mRNA. However, the proportion of thymidine is high, perhaps indicating the presence of some poly(T) sequences in the product complementary to the poly(A) sequences in the mRNA.
Such poly(T) sequences would presumably interfere with the specificity of the DNA product, since poly(A) sequences are common to many eukaryotic mRNAs (15,(25)(26)(27).
To prevent these putative poly(A)-poly(T) sequences from being scored in assays for secondary structure, the DNA product in all subsequent experiments was synthesized with deoxycytidine and deoxyguanosine as the only labeled nucleotides, and X1 nuclease was used to analyze the secondary structure of nucleic acids.
To show that these precautions are sufficient to prevent poly(A)-poly(T) duplexes from being scored, the experiment shown in Table III   [3H]DNA product at 1 ng per ml was annealed for 1 hour either to itself or to poly(A) at 50 rg per ml to a C,t of 1 mo1e.s per liter relative to the poly(A).
[3H]Poly(T) (see "Materials and Methods") was annealed to a 1000.fold excess of poly(A) to a C,t value of 1 mo1e.s per liter.
The samples were then assayed for secondary structure by S1 nuclease. or total oviduct RNA with the DNA product are illustrated in Fig. 2. The reciprocal of the percentage nuclease resistance was plotted as a function of the reciprocal of the C,t (see "Materials and Methods"). For both ovalbumin mRNA and total oviduct RNA, the data (Fig. 2)  [IH] DNA product ( 5 ng per ml) was annealed with purified ovalbumin messenger RNA (A), prepared as described under "Materials and Methods," or with total hen oviduct RNA (B), prepared as described in the legend to Table IV. After annealing, samples were assayed for S1 nuclease resistance as described under "Materials and Methods."

Conditions
The concentration of RNA in the experiment shown in A was either 100 (o), 170 (A), or 900 (0) ng per ml. In the experiment shown in B, all samples were annealed for 2 hours, and the RNA concentration varied from 100 ng per ml to 200 pg per ml.
The C,t values were based on the concentration of the purified ovalbumin mRNA or hen oviduct RNA.
species found in both total hen oviduct RNA and our purified mRNA.
The alternate interpretation, that the DNA product is complementary to multiple species of RNA, would require that each of those RNA species be present at nearly identical concentrations (see "Discussion"). The C,tliz for the hybridization reaction with purified ovalbumin mRNA is 3.8 X low3 mole's per liter. To gain further information on the RNA complementary to the DNA product, the experiment described in Fig. 3 Fig. 3 shows the results of the protein synthesis assay.
For each RNA sample, ovalbumin synthesis is proportional to RNA input.
The immunoadsorbed polysomes have a higher specific ovalbumin-synthesizing activity than total polysomes, and the nonimmunoadsorbed polysomes a lower activity, as we have previously shown (8). Panel B of Fig. 3  Genome-To measure the absolute number of copies of the sequence coding for ovalbumin in the chicken genome, the following experiment was performed. A large amount of either chicken liver or chicken oviduct DNA was denatured and reassociated in the presence of [3H]DNA product specific for ovalbumin and a trace amount of 14C-labeled unique sequence chicken DNA.
If the ratio of sequences in the unlabeled cellular DNA to identical sequences in the labeled DNA is high, the rate of reassociation of the labeled sequences will be determined by the concentration of identical unlabeled cellular sequences. Therefore by comparing the Cotm of the labeled, ovalbumin-specific product to that of the labeled unique sequence DNA, the relative reiteration of these two species in the genome can be determined.
From the data presented in Fig. 4 it is clear that for both liver DNA (Panel A) and oviduct DNA (Panel B) the C&z of the ovalbumin sequences is not significantly different from that of the unique sequences.5 From these results we conclude that both

B=D
The single stranded DNA product and the double stranded, cellular DNA were melted and mixed together (Fig. 4). All DNA of the same sequence should now react at the same rate, regardless of whether it originally was from the DNA product or the cellular DNA.
Since 707, of the labeled DNA reassociated at the maximum Cct value t,ested: DNA (all at 5 mg per ml), denatured, and reannealed as described in under "Materials and Methods." The purpose of the salmon sperm DNA was to control for the effect of viscosity on the rate of the reassociation reaction. Pilot experiments indicated that the DNA product does not react with salmon sperm DNA.
At different times aliquots were taken and assayed for secondary structure by the use of S1 nuclease as described under "Materials and Methods." in oviduct and in liver there are 2 copies of the ovalbumin sequence per diploid genome.
The relative reiteration of ovalbumin sequences in oviduct and non-oviduct tissue can also be measured using double stranded DNA and the technique of Gelb et al. (29). In this procedure double stranded DNA product is mixed with a constant amount of either oviduct or liver DNA. The presence of ovalbumin sequences in the cellular DNA will cause the DNA product to reassociate more rapidly than it would alone. If the number of ovalbumin sequences per unit mass of DNA is the same in both oviduct and liver DNA, the rate of reassociation of double stranded DNA product in the presence of a given amount of either of these DNAs will be the same. The data shown in There are a minimum of 2.3 copies of the ovalbumin sequence present in the cellular DNA for every 1 in the DNA product.
If the reaction proceeds further at greater Cot values, this ratio will be even greater.
The maximum possible effect of the DNA product on the CJt,z of the cellular DNA, and on our estimate of the number of ovalbumin sequences per diploid genome, is to increase it by a factor of ( and these molecules were then denatured and assayed for secondary structure (Fig. 6). In order to have a control of perfectly paired DNA duplexes specific for ovalbumin, double stranded DNA product was reassociated only with itself, then denatured and assayed under identical conditions. We conclude that a very high percentage of the nucleotides in the cellular DNA-DNA product duplex are correctly base paired.

DISCUSSION
Analysis of the data presented in this paper depends on the homogeneity and specificity of the RNA template used in the synthesis of the DNA product.
We have previously presented evidence concerning the purity of our ovalbumin mRNA (8, 10). In brief, ovalbumin polysomes are selectively isolated from the total oviduct polysomes by immunoadsorption with antiovalbumin antibody and an ovalbumin matrix (see "Materials and Methods").
A variety of indirect controls with polysomes from other tissues, with other antibodies and with other matrices, indicated that both the binding of antibody to the polysomes and the subsequent binding of the polysome-antibody complex to the matrix have a high degree of specificity (8, 10). In more direct controls, the pattern of nascent peptide chains in the immunoadsorbed polysomes was studied and found to be consistent with that expected for pure ovalbumin polysomes (8 is purified from other oviduct RNAs by immunoadsorption in an ovalbumin-anti-ovalbumin system to the same degree that ovalbumin mRNA is purified (Fig. 3). The size of the DNA product (7 S, Fig. 1 (15), and mRNA for mouse k chain immunoglobulin (30), have all been approximately the same size as the product reported here.
However, the short chain length of the DNA does not preclude extensive or complete representation of template nucleotide sequences in the enzymatic product (33). In Fig. 2 the data for hybridization of DNA product with ovalbumin mRNA fit well to a straight line. We conclude that almost all of the DNA product is complementary to one homogeneous RNA species.
The alternate interpretation, that the DNA product is complementary to multiple species all of which hybridize at the same rate with complementary sequences in the DNA product, seems unlikely, since it would require each of those multiple RNA sequences to be present at the identical molar concentration, each to be specific to the oviduct (Table  IV), and each to be purified to the same degree by immunoadsorption (Fig. 2).
The presence of minor contaminating DNA species in the DNA product, complementary to sequences other than that of ovalbumin mRNA, would not invalidate the conclusions drawn from the experiments presented here since the experiments are based on the kinetics of hybridization of the major kinetic component of the DNA product.
We do not have a sufficient amount of the DNA product to measure its concentration optically. We can only estimate the concentration of DNA product from the amount of radioactivity present and the specific activity of the DNA. Therefore we can neither get a reliable estimate of the complexity of the double stranded DNA, nor can we calculate the absolute number of copies of the ovalbumin gene from Fig. 5. Ovalbumin comprises approximately 60 to 65% of the soluble protein synthesized in the oviduct (lo), yet the number of ovalbumin sequences seems to be 1 per haploid chicken genome, and seems not to be amplified in tissue specialized for ovalbumin synthesis.
Similar evidence exists for a lack of gene amplification in hemoglobin-specific DNA sequences in duck (5, 7) and mouse (6) reticulocytes, and in silk fibroin-specific sequences in Bombyx mori (34). At least in these cases the considerable selective amplification of a portion of the information of the cell genome occurs at some level other than that of the number of gene copies in the cell.
In previous studies on the oviduct, ovalbumin-specific nucleic acids have been operationally defined as nucleic acids capable of synthesizing ovalbumin in vitro (18,35). This criterion has the advantage of being very stringent, particularly if tryptic digests of the polypeptide are analyzed (18), but has the disadvantage that the sequence coding for the ovalbumin polypeptide will not be detected if the polynucleotide containing that information is not in the form translatable by ribosomes.
In this work we have developed a quantitative assay for the ovalbumin sequences which is based on a different criterion, namely, rate of hybridization with a specific DNA sequence.
This method has the advantage that the sequence can still be detected in nucleic acids which are not translated by ribosomes. The stringency of this method is dependent on the stringency of the assay for secondary structure of the hybrids. Denaturation of the hybrids, our most stringent assay, can detect perhaps 3% mismatching by a 2' decrease in T, (36). This hybridization assay should prove a useful method for quantitation of ovalbumin sequences in different physiological or developmental states and should be useful in analysis of the structure of the genome in higher organisms.