Purification of ovalbumin messenger ribonucleic acid by specific immunoadsorption of ovalbumin-synthesizing polysomes and millipore partition of ribonucleic acid.

Abstract Ovalbumin-synthesizing polysomes were isolated from total oviduct polysomes by reaction with anti-ovalbumin antibody and adsorption to a matrix of glutaraldehyde-linked ovalbumin. The RNA from immunoadsorbed ovalbumin polysomes was then deproteinized and selectively adsorbed to Millipore filters. Several types of experiments indicated the resulting RNA fraction is highly enriched for ovalbumin mRNA.

polysomes were isolated from total oviduct polysomes by reaction with anti-ovalbumin antibody and adsorption to a matrix of glutaraldehyde-linked ovalbumin.
The RNA from immunoadsorbed ovalbumin polysomes was then deproteinized and selectively adsorbed to Millipore filters. Several types of experiments indicated the resulting RNA fraction is highly enriched for ovalbumin mRNA.
As an approach to understanding the molecular events involved in the regulation of specific protein synthesis in higher organisms, several laboratories have focused on the isolation of mRNA specific for a given protein.
The mRNAs for hemoglobin (l-5), myosin (6,7), the light chain of an immunoglobulin (8,9), ovalbumin (lo), lens a-crystallines (11, la), histones (13)(14)(15)(16), and silk fibroin (17) have been identified and in some cases purified and partially characterized. The methodology followed for the isolation of specific mR.NA in most cases has been based on the fact that the protein comprises a large percentage of the total protein synthesized in the system, in addition to a special characteristic of the protein such as size or a unique amino acid composition.
This type of methodology restricts the possibility of isolation of specific mRNAs to a very few systems. An immunological approach based on the specificity of an antibody prepared against the native protein and reactive against nascent polypeptide chains associated with mRNA in polysomes has been used in our laboratory to purify ovalbumin mRNA from other messengers. This approach has the possible advantage of being b vencrally applicable to the isolation of specific mRNAs.
We have demonstrated that [1251]anti-ovalbumin binds specifically to ovalbumiusynthesizing polysomes (30) and that such polysomes can be precipitated by sequential addition of ovalbumin and anti-ovalbumin (30,31). Ovalbumin nascent chains and ovnlbumin mRNA activity have been shown in the immunoprecipitate (31). When attempting to isolate large amounts of polysomes, this precipitation technique has some disadvantages: a large excess of antibody must be used, and the polysomal components must be extracted from the precipitate that contains a large amount of protein.
This excess protein increases the possibility of RNA degradation.
In the present study, as a first step in the purification of ovalbumin mRNA, ovalbumin-synthesizing polysomes were isolated by a method based on immunoadsorption. This technique avoids the disadvantages of immunoprecipitation.
The second phase in the purification of ovalbumin mRNA was based on a difference in physicochemical properties between mRNA and rRNA.
Several workers have purified presumptive mRNA on the basis of an adenine-rich sequence attached to the message. l?oly(dT) cellulose (32) and poly(U) cellulose (33,34) have been used to isolate presumptive messengers containing an adenine-rich region from heterogeneous RNA populations. Selective adsorption to Millipore has been reported by Brawerman et al. (35) to enrich for mRNA in two different types of cells and used by Phillipson et al. (33) to isolate RNA fractions containing an adenine run from virus-specific RNA. During the course of this work Means et al. (36) showed that hen oviduct rapidly labeled poiysomal RNA isolated from a sucrose gradient region that contains ovalbumin mRNh activity (8 to  and preparation of polysomes were as described previously (30) but using 500 fig per ml of heparin in the homogenization buffer.
Preparation of Anti-OV y-Globulin Fraction-Goats were immunized with electrophoretically homogeneous ovalbumin purified as previously described (38). Five milligrams of ovalbumin, mixed with complet,e Freund's adjuvant, were injected intramuscularly.
After 3 weeks a second injection of the same mixture was made, and 2 weeks later goats were bled through the carotid artery.
Serum was separated and precipitated twice with 40% saturated ammonium sulfate. The y-globulin fraction obtained was dialyzed against 0.01 M sodium phosphate buffer, pH 7.2, containing 0.015 M NaCl, and passed through a column (5 cm in diameter) containing 10 cm of CM-cellulose over 10 cm of DEAE-cellulose, equilibrated with the same buffer. This column removes RNase activity (30).
Preparation of Ovalbzonin, BSA, rind Anti-OV Matrices-The proteins were covnleutly cross-linked with glutaraldehyde according to the method of Avrameas and Ternynck (39). Commercial ovalbumin or USA was dissolved in 0.1 M sodium phosphat.e buffer, pH 7.0, to a final concentration of 50 mg per ml. For each milliliter of protein solution, 0.1 ml of 12.5y0 glutaraldehyde was added dropwise.
The mixture was allowed to gel for 4 hours at room temperature, homogenized three times in 0.2 M sodium phosphate buffer, pH 7.3, and then washed three times with 0.01 M sodium phosphate buffer, pH 7.3, containing 0.15 Y sodium chloride.
When the gels were used for isolating polysomes, they were then washed three times with polysome buffer (0.025 M Tris-HCl, 0.025 M NaCl, 0.005 1c1 MgClg, pH 7.6, at 4") containing 100 pg of heparin per ml.
The y-globulin fraction isolated from goat serum was used to prepare the anti-ovalbumin matrix.
To each milliliter of the y-globulin preparation (45 mg per ml) 0.1 ml of 1.0 M sodium phosphate buffer, ~1-1 7.0, and 0.05 ml of 12.57; glutaraldehydc were added. The misture was allowed to gel for 30 min at room temperature and homogenized and washed as described above. All gels were stored at 4" for up to 1 week before using. OV-mat.rix, ovdbumin matrix. and the supernatant discarded.
The matrix was washed with 10 ml of 0.01 M sodium phosphate buffer, pH 7.3, 0.15 M NaCl, centrifuged, and the supernatant discarded.
The washing was repeated three more times. The adsorbed protein was eluted with 0.1 M glycine-HCl buffer, pH 2.8, as described by Avrameas and Ternynck (39). The matrix was continually mixed for 10 min on a magnetic stirrer at room temperature with 2 ml of the glycine buffer. The mixture was then centrifuged 10 min at 6000 rpm and the supernatant saved. The elution was repeated two more times, and the supernatants were pooled and dialyzed overnight against 0.15 M NaCl, 0.01 M sodium phosphate buffer, pH 7.3. The material obtained comigrated as a single peak in an SDS-acrylamide gel with a [3H]ovalbumin standard prepared as described by Palmiter et al. (38). The [14C]ovalbumin was used as a marker in the experiments reported in this paper.
Purification of Anti-08 Antibody with Ovalbumin Matrix-The anti-ovalbumin y-globulin fraction isolated from goat serum (see above) was incubated with an ovalbumin matrix (20 ml of y-globulin to 1 g wet weight of matrix) for 45 min at room temperature with continuous stirring.
The mixture was centrifuged, washed four times with 0.15 M NaCl, 0.01 M sodium phosphate buffer, pH 7.3, and eluted with 0.1 M glycine-HCl buffer, pH 2.8, as described above. The sample was dialyzed overnight against 0.15 M NaCl, 0.01 M sodium phosphate buffer, pH 7.3, and frozen in small aliquots.
The titer of y-globulin preparation was increased approximately 7-fold after purification. One milligram of the purified preparation precipitated 100 pg of ovalbumin at the equivalence point.
The antibody was more than 90% pure as indicated by the amount of y-globulin precipitated at the equivalence point.
In all the experiments reported in this paper the pure anti-ovalbumin antibody was used. Isolation of Ovalbumin-synthesizing Polysomes-Polysome preparations at a concentration of 10 A260 units per ml in polysome buffer (see above) containing 100 pg per ml of heparin were incubated with pure anti-OV (1 mg of antibody per 20 A260 units of polysomes) at 4' for 45 min. The reaction mixture was transferred to a Cores centrifuge tube containing ovalbumin matrix (400 mg/20 A260 units of polysomes) and incubated with constant stirring at 4' for 45 min. The preparation was centrifuged in a Sorvall HB-4 swinging rotor for 10 min at 6000 rpm and the supernatant saved. The matrix was washed with 0.5 M sucrose, 0.15 M NaCl, 1 7. Triton X-100, 1 y. DOC in polysome buffer with 100 /Ig per ml of heparin (4 ml of buffer per 400 mg wet weight of matrix), centrifuged as before, and the supernatant saved in a separate tube. This washing procedure was repeated twice more. The final washing was with polysome buffer alone in order to remove the detergents and sucrose. To elute the adsorbed polysomes, a buffer containing 0.01 M Tris-HCl, 0.05 M EDT-%, pH 7.5 and 100 pg of heparin per ml was added (2 mI/400 mg wet weight of matrix) and incubated 15 min at 4" with constant stirring.
The preparation was centrifuged as before and the supernatant saved. Two more elutions were performed and each of the supernatants saved in separate tubes. All fractions collected were made 0.05 M in EDTA, 0.15 M in NaCl, and 1% in SDS. The fractions were ethanol-precipitated overnight as described below.
Ethanol Precipitation of RNA-RNA samples were made 0.1 M in NaCl and adjusted, if necessary, to neutral pH by addition of 0.2 volume of 1 M Tris-HCI, pH 7.0. Two volumes of ethanol were added and the RNA precipitated a minimum of 6 hours at -20".

The precipitated
RNA was pelleted out of the ethanol at 14,000 X g for 20 min at 0".

542
Deproteinization of RNA-Two methods of deproteinizing RNA from polysomes were used alternately or in combination as indicated.
In one method polysomes were made 0.05 M in EDTA, 1% in SDS, and 0.1 M in NaCI, and precipitated with 2 volumes of ethanol as described above for RNA.
The precipitate was dissolved in 1 ml of 0.5% SDS-acetate EDTA buffer (0.02 M sodium acetate, 0.005 M EDTA, pH 5.0). The sample was then layered on a 11.5~ml 5 to 20% sucrose gradient in 0.5% SDS-acetate EDTA 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 the material in the lower part of the gradient (including all the 18 S RNA peak) was collected and precipitated with ethanol as described above. The material in the upper part of the gradient, consisting of the tRNA, SDS-treated proteins, and heparin (added as a nuclease inhibitor during polysome isolation) was discarded.
Alternatively polysomes were made 1% with Sarkosyl, salt and ethanol added as indicated above, precipitated at least 6 hours at -2O", and the precipitate dissolved in 1% Sarkosyl, 0.01 M EDTA, Tris-acetate, pH 6.5. Dry heat-sterilized CsCl was then added to this solution to a density of 1.723 to 1.739 g per cc (about 105 g of CsCl per 80 ml of solution).
Aliquots of the sample (8.8 ml) were dispensed into Beckman polyallomer tubes, overlayered with mineral oil, and centrifuged 60 hours at 21", 33,000 rpm, in a Beckman type 40 rotor.
Under these conditions the larger RNA species pelleted while the detergenttreated proteins, any DNA present, and tRNA, which does not come to equilibrium because of its small size, remained in the CsCl solution.
At the end of the centrifugation, oil and water were drawn off and the RNA pellet redissolved in 0.01 M Tris-Cl, pH 7.5, and ethanol-precipitated several times. Up to 10 mg of RNA can be loaded on one such gradient.
We find the SDS-sucrose gradient technique results in RNA with a higher specific protein-synthesizing activity, while the CsCl technique is more suited to bulk deproteinization of RNA. Reticutocyte &sate Assay-Ovalbumin synthesis was measured by immunological precipitation of ovalbumin from a rabbit reticulocyte lysate protein-synthesizing system. This procedure was as described by Rhoads et al. (lo), except that the antibody precipitates were washed by pelleting through sucrose as described by Rhoads and that in some reactions, where indicated, 3H-labeled leucine was replaced with [3H]isoleucine, specific activity 6 Ci per mM, at 24 PCi per ml of reaction mixture.
Polyacrylamide Gel Electrophoresis-For RNA this procedure was performed as described by Maize1 (40). The "Neutral SDS-EDTA" system was used, and gels were pre-electrophoretitally treated I hour at 5 ma per gel. Gels were scanned at 260 nm in a Gilford recording spectrophotometer with a linear transport device. For protein the procedure was as described by Palmiter et al. (38).
The filter was placed in a Millipore fritted glass collection apparatus and washed 10 times under vacuum with 10 ml of Buffer A. 3 The RNA sample, in Buffer A, was allowed to flow through the filter by gravity.
Up to 10 mg of RNA in 100 ml of buffer have been passed through one 2 Robert E. Rhoads, manuscript in preparation. filter.
After the sample had passed through, the filter was washed 10 times with lo-ml aliquots of Buffer A under vacuum. This procedure was done at 0 and 20" with no apparent differences in the RNA fractions obtained.
The filter was then removed and placed in a sterile, covered Petri dish. Three milliliters of Buffer A were added, the dish agitated, briefly, and the buffer removed and discarded.
This step was to remove any RNA trapped on the filter under the edges of the glass chimney. Three milliliters of low salt buffer (0.01 M EDTA, 0.5% SDS, pH 5.0) were added, the filter agitated for 15 min at 20", and the solution removed and saved. This procedure was repeated four more times, and the combined fractions were ethanol-precipitated three times.
Low pH buffer was used to diminish the possibility of RNA degradation.

Isolation of Ovalbumin-synthesizing
Polysomes-The method developed for isolating ovalbumin-synthesizing polysomes involves: (a) incubation of oviduct polysomes with anti-ovalbumin which binds specifically to ovalbumin-synthesizing polysomes (30) ; (b) reaction of the antibody-polysome complex with immobilized ovalbumin prepared by cross-linking with glutaraldehyde; (c) washing of the matrix to remove nonspecifically trapped material; and (d) elution of the bound polysomes with EDTA which liberates the ribosomal subunits and mRNA from the nascent polypeptide chains that remain attached to the y-globulin-ovalbumin gel complex. As seen in Fig. 1, when hen oviduct polysomes were treated in this way, some AzeO absorbing material was bound to the gel and released with EDTA.
That the adsorption of polysomes to the OV-matrix depends on its previous reaction with anti-OV is shown in Fig. 2. When poly- 1. Immunoadsorption of ovalbumin-synthesizing polysomes. Hen oviduct polysomes (100 A260 units) were incubated with anti-ovalbumin and reacted with ovalbumin matrix as described under "Materials and Methods." The preparation was centrifuged, the supernatant was saved (Fraction 1). and the adsorbant was washed three times with polvsome buffer (Fractions 2 to 4). Elution with EDTA was as desciibed under "Materials and Methods" (shaded fractions). Each fraction was made 0.05 M in EDTA, 1% in SDS, and 0.1 M in NaCl and precipitated overnight at -20" with 2 volumes of ethanol.
The samples were centrifuged 15 min at 10,000 rpm, the precipitate was dissolved in 0.570 SDS-acetate EDTA buffer (see "Materials and Methods"), and adsorbance at 260 nm was measured. The 'eluted fractions were precipitated with ethanol and their A260 measured as described in Fig. 1.
somes were not incubated with antibody, the amount of A~,x absorbing material eluted from the ovalbumin matrix was very small (less than 2%). The results obtained with different concentrations of antibody indicate that to attain saturation a very large amount of anti-OV would be required. This is in accordance with the binding of anti-OV to hen oviduct polysomes shown previously (31). Control experiments indicated that at 3 mg of anti-OV per 20 Am units of polysomes the amount of ovalbumin matrix used was not limiting for adsorption.
In the remainder of this study we have used a concentration of 1 mg of anti-OV per 20 AHO units of polysomes, a concentration which results in isolation of 15 to 20% of the total hen oviduct polysomes.
The specificity of this method for isolating ovalbumin polysomes was determined by the following experiments.
In the experiment shown in Fig. 3 hen oviduct polysomes were incubated with anti-BSA and then with a BSA matrix.
When the matrix was washed only with polysome buffer approximately 8% of the AseO absorbing material appeared upon elution with EDTA, indicating a large amount of nonspecific binding. However, washing the matrix with sucrose, detergents, and salts (see "Materials and Methods") prior to EDTA elution reduced this nonspecific trapping to less than 2%. This result is in accordance with that reported for the precipitation of ovalbumin-synthesizing polysomes (31) and indicates the usefulness of the detergent-sucrose-washing procedure.
It is interesting that the washing with sucrose, detergents, and salt did not diminish the amount of material eluted when anti-OV and ovalbumin matrix were used with hen oviduct polysomes.
This indicates that the amount of nonspecific trapping depends on the antigen and antibody used, and that appropriate control experiments must be done if these techniques are to be applied to other systems.
In the rest of the experiments presented the immunoadsorbents were washed with sucrose, detergents, and salt in addition to the polysome buffer.
To quantify the specificity of the method for isolating ovalbumin polysomes the experiment presented in Fig. 4 was performed.
The specific radioactivity of the liver polysomes was such that 80% of the AzcO absorbing material was from oviduct polysomes.
The mixed polysomes were incubated with anti-OV and ovalbumin matrix, washed and eluted as described above. The eluted material contained 19% of the Am absorbing material and 2% of the radioactivity. This result indicates that the contamination of the eluted sample by nonspecifically trapped polysomes was about 10% of the total adsorbed polysomes. Assuming the starting material contains 60% ovalbumin-synthesizing polysomes and that nonspecific trapping is random, the polysome preparation after immunoadsorption is estimated to be more than 95% pure.
To prove directly the specificity of the immunoadsorption method ovalbumin mRNA was measured by its capacity to direct ovalbumin synthesis in a heterologous cell-free system. Under RNase-free conditions, the specific ovalbumin-synthesizing activity (ovalbumin synthesized per amount of RNA) of the RNA extracted from immunoadsorbed polysomes should be higher than that from total polysomes.
Such an experiment is presented in Fig. 5. Hen oviduct polysomes were treated as described for the isolation of ovalbumin-synthesizing polysomes. RNA was extracted from different fractions and its capacity to synthesize ovalbumin in vitro was measured.
The specific synthesizing activity of the RNA from adsorbed polysomes was higher than that of the original polysomes while that of the nonadsorbed polysomes was lower.
That RNase activity was essentially absent during the treatment of polysomes is indicated by the fact that the specific ovalbumin-synthesizing activity of the RNA was not decreased by incubating the polysomes with The presence of ovalbumin mRNB activity in the nonadsorbed polysomes can be explained by the fact that the concentration of anti-OV used was less than saturating.
The increase in specific synthesizing activity after purification of the ovalbumin polysomes was on the order of the expected increase. Ova.lbumin represents 60 to 65% of the total protein synthesized in the hen oviduct.
Thus assuming the starting polysomes are also 60 to 65% specific for ovalbumin, the highest purification expected would be only about 1.5-to 1.6-fold.
To show more clearly the purification of ovalbumin-synthesizing polysomes, the experiment shown in Fig. 6 was performed. Immature chicks stimulated with estrogen were used instead of adult hens. Chicks received a secondary stimulation with estrogen and 18 hours later polysomes were prepared from the oviduct magnum.
The relative rate of ovalbumin synthesis was measured by the technique of Palmiter et al. (38) in explants of the tissue in culture and found to be 177,. Therefore, the highest purification expected would now be about 6-fold.
Oval bumin-synthesizing polysomes were purified by the immunoadsorbent method and the specific ovalbumin-synthesizing activity of the RNA determined. Approximately 9.5% of the &so absorbing material was adsorbed, then eluted from the matrix.
This material showed a 7-fold increase in specific ovalbumin-synthesizing activity compared to the original polysomes, while the nonadsorbed polysomes exhibited a 4-fold decrease in specific ovalbumin-synthesizing activity. Additional proof of the specificity of the isolation of ovalbumin polysomes was obtained by comparison of the nascent polypeptides associated with the immunadsorbed polysomes with those of the total polysomes (Fig. 7). Nascent chains from adult hen oviduct were labeled with 5-min pulse of Wlabeled amino acids in vitro. We have previously shown that with a puIse of this duration nascent polypeptides and not ribosomal proteins are labeled (31). Polysomes isolated from the tissue were treated with anti-OV and OV-matrix and eluted with 1% SDS. The pattern of nascent chains was determined by acrylamide-SDS electrophoresis.
Immunoadsorbed nascent polypeptides showed a pattern in accordance with that expected for ovalbumin nascent chains, i.e. there was a sharp increase in radioactivity in the region where the ovalbumin standard appeared and then a gradual decrease. This pattern is significantly different from that seen in Fig. 7 (Table  I).
For both the messenger activity and for the poly(A), a fraction was bound to the filter and a fraction was not. Repassing each of these two fractions showed that the fraction which was bound on the first pass was predominantly bound on the second, and that very little of the material which did not bind on the first pass bound on the second. Bound and unbound RNB were then compared with total polysoma.1 RNA in specific protein-synthesizing activity. Each of these RNA fractions was assayed in the reticulocyte lysate system at three RNA concentrations for its ability to synthesize ovalbumin.
The results, shown in Fig. 8 tein synthesizing activity: the percentage of mRNA molecules could be higher in the bound fraction than in the total RNA; the RNA in the bound fraction could in some way be activated and be a more efficient template for protein synthesis; or there could be an inhibitor of protein synthesis in the total RNA, which was removed from the bound fraction by Millipore treatment. If the first of these three possibilities is the cause of the difference in specific synthesizing activity, the profile on polyacrylamide gels of the bound RNA would be expected to be different from that of total polysomal RNA. Fig. 9 shows the AZGO profiles of two gels, each with the same amount of either total polysomal or Millipore-bound RNA.
The rRNA peaks appear relatively reduced in the Millipore-treated fraction, and RNA species migrating mainly near the 18 S peak are now apparent.
Since in this experiment total polysomal RNA was used as the starting material, the nonribosomal RNA species probably represent the pattern of mRNAs from the hen oviduct. Competition hybridization of mRNA for ribosomal genes by the method of Brown and Weber (41) also indicates a reduction of rRNA in the Millipore-bound fraction.
Rough estimates indicate that approximately two-thirds of the RNA in this fraction is ribosomal RNA. To test whether the sedimentation profile of ovalbumin-synthesizing activity had been changed by Millipore binding of the RNA, the following experiment was performed. The RNA fraction which bound to Millipore was centrifuged on a 5 to 20% sucrose gradient, and the fractions assayed for ovalbumin messenger activity.

546
The results shown in Fig. 10 metrical peak slightly more slowly than 18 S rRNA. This result agrees with that described by Rhoads et al. (10) for ovalbumin messenger activity in total oviduct polysomal RNA.
Millipore Partition of RNA Extracted from ImmunopuriJied &albumin Polysomes--In order to obtain a highly purified preparation of ovalbumin mRNA, both immunoadsorption of ovalbumin polysomes and Millipore partition of RNA were combined in a single experiment (Fig. 11). Hen oviduct polysomes (600 AzGO units) were adsorbed and eluted from OV-matrix, and a fraction of both the adsorbed and the nonadsorbed polysomes was passed through Millipore filters. To test for the purification of ovalbumin mRNA, different fractions were assayed for ovalbumin-synthesizing activity in the reticulocyte lysate system. As shown in Fig. 11, the specific synthesizing activity increases with immunoadsorption of polysomes and with Millipore partition of RNA.
This indicates that the two techniques can be used in series, to provide an RNA fraction greatly enriched for a specific mRNA. DISCUSSION In order to purify a specific mRNA two different problems must be faced: (a) the purification of the specific mRNA from Oviduct polysomal RNA was deproteinized by the C&l technique and passed over a Millipore filter, and the bound fraction was centrifuged on a 5 to 20y0 sucrose gradient (see "Materials and Methods"). Fractions were collected, ethanol was precipitated three t,imes, and the RNA of each fraction was brought IIP in 200 ~1 of HzO. Aliquots of each fraction (10, 25, and 50.~1) were assayed in the reticulocyte lysate protein-synt,hesizing system using leucine as the labeled amino acid.
The value plot&ed represents the counts per min of ovalbumin synt.hesized in t,he assay of the 25-111 aliquot. For all fractions the relation between amount of RNA added and ovalbumin synthesized was nearly linear. To provide a size marker total oviduct polysomal RNA was sedimented in another tube and collected through a Gilford recording spectrophotometer with a 2-mm light path flow cell. Absorbance at 260 nm was measured.
other messengers and (b) the purification of the mRNA from other types of RNA, mainly rRNA. Different techniques must be applied to each of these problems.
The purification of ovalbumin mRNA from other messengers has been based on the specific immunoadsorption of ovalbuminsynthesizing polysomes.
The specificity of the immunoadsorption technique is evidenced by the following: (a) polysomes that have not been reacted with anti-OV do not bind to the ovalbumin mabrix; (b) increasing the amount of the antibody increases the adsorption of the polysomes (Fig. 2) ; (c) hen oviduct polysomcs do not react when the immunoadsorbent system is formed by a protein (BSA) not found in the oviduct and its corresponding ant,ibody (Fig. 3); (d) polysomes from a tissue that does not synthesize ovalbumin (liver) do not react with the anti-OV and ovalbumin system (Fig. 4) ; (e) the capacity of the RNA to synthesize ovalbumin in U&O is increased in immunoadsorbed polysomes (Figs. 5 and 6) ; (f) nascent polypeptide chains isolated from immunoadsorbed polysomes show a pattern that is in accorda.nce with that expected for ovalbumin nascent chains (Fig. 7).
Other methods for isolating specific mRNA have been based on the sedimentation characteristics of the RNA (1,2,7,8,11,12,17). Immunological techniques have the potential of being generally applicable and are not restricted to systems in which a protein comprises a very large percentage of the total protein being synthesized.
As illustrated by the experiment in Fig. 6, ovalbumin-synthesizing polysomes were specifically immunoadsorbed when oviduct-synthesizing 177, ovalbumin was used. By comparing the capacity to synthesize ovalbumin in vitro of the RNA extracted from immunoadsorbed and total polysomes it can be concluded that a high degree of purification was anti-OV and OV-mat.rix, washed, and elut.ed with EDTA. bumin in vitro (Fig. 8). The electrophoresis pattern of RNA indicated that the relative concentration of rRNA had decreased (Fig. 9). By rough calculation mRNA should comprise approximately 1% by mass of polysomal RNB.
Therefore, an increase of 26-fold in specific protein-synthesizing activity suggests that the fraction after Millipore treatment contains about 25% mRNA.
We have also been able to isolate an RNA fraction with increased specific ovalbumin-synthesizing activity by the phenol partition method of Smith et al. (43). We prefer the Millipore method because of the difficulty in quantitative recovery of messenger activity from the phenol phase, and because exposure to phenol induces an aggregation of ovalbumin mRNA molecules which becomes apparent upon rate sonal centrifugation in aqueous solution. 5 Such aggregation after exposure to phenol has been reported for other RNA species (44).
Our results indicate ( Table I) that some ovalbumin mRNA binds to Millipore filters and some does not. We feel one explanation for this behavior could be a difference in the length, or in the presence, of the adenine-rich sequence presumed to be on ovalbumin mRNA.
Whether this heterogeneity exists normally in the cell or is an artifact of RNA isolation remains to be seen.
We have evidence that our polysome preparations contain little or no nonpolysomal RNA (30, 31). Both deproteinizing procedures used in our laboratory remove the 4 S and 5 S RNAs, so the major RNA species remaining after immunoadsorption of polysomes and Millipore treatment of RNA should be ovalbumin mRNA, 18 S rRNA, and 28 S rRNA.
Since ovalbumin mR.NA activity sediments in a sucrose gradient near the 18 S region before (10) or after Millipore treatment (Fig. lo), rate zonal sedimentation of the Millipore-treated RNA should separate ovalbumin messenger from the 28 S rRNL4 contaminant. This added step should result in an RNA preparation purer for ovalbumin messenger. Our immediate interest in immunologically purified mRNA is to synthesize complementary DNA sequences with Rous Sarcoma Virus DNA polymerase (45). Under our conditions for DNA synthesis, rRN,4 is not significantly transcribed.6 Therefore, the amount of rRN-4 present in our purified messenger preparations should not interfere with the synthesis of a homogenous hybridization probe complementary only to ovalbumin mRNA.
Bcknowledg?ner&s-We wish to espress our thanks to Ruy Perez Tamayo and J. Michael Bishop for their encouragement and their many helpful suggestions during the course of this work, and to R. Rhoads, F. Dice, J. Taylor, and D. Shapiro for their critical reading of this manuscript.