Isolation of Hen Oviduct Ovalbumin and Rat Liver Albumin Polysomes by Indirect Immunoprecipitation*

SUMMARY The polyribosomes synthesizing ovalbumin and albumin have been isolated from total oviduct polysomes and total rat liver polysomes, respectively. The isolation was performed using an indirect immunoprecipitation technique which is based on the specific reaction which occurs between antibody against a purified native protein and the nascent peptide chains on polysomes synthesizing that protein. A soluble antibody-nascent chain-polysome complex is formed by incu-bating antibody with polysomes, and then precipitated by reaction with an anti-antibody. The techniques developed are both efficient and highly specific. Near quantitative isolation of polysomal mRNA coding for a specific protein may be achieved. Indirect immunoprecipitation results in nonspecific precipitation of less than 0.5 % of total polysomes. Ovalbumin mRNA isolated by indirect immunoprecipitation is 99% pure with respect to contamination by other species of mRNA and rat liver albumin mRNA is greater than 95% pure.


SUMMARY
The polyribosomes synthesizing ovalbumin and albumin have been isolated from total oviduct polysomes and total rat liver polysomes, respectively.
The isolation was performed using an indirect immunoprecipitation technique which is based on the specific reaction which occurs between antibody against a purified native protein and the nascent peptide chains on polysomes synthesizing that protein.
A soluble antibody-nascent chain-polysome complex is formed by incubating antibody with polysomes, and then precipitated by reaction with an anti-antibody.
The techniques developed are both efficient and highly specific. Near quantitative isolation of polysomal mRNA coding for a specific protein may be achieved.
Indirect immunoprecipitation results in nonspecific precipitation of less than 0.5 % of total polysomes. Ovalbumin mRNA isolated by indirect immunoprecipitation is 99% pure with respect to contamination by other species of mRNA and rat liver albumin mRNA is greater than 95% pure.
Evidence supporting these conclusions includes: (a) incubation of oviduct polysomes containing labeled nascent peptide chains with anti-bovine serum albumin results in precipitation of only 0.4% of the labeled polysomes; (b) indirect immunoprecipitates from oviduct polysomes reacted with anti-ovalbumin contain essentially no nascent peptide chains larger than ovalbumin; (c) ovalbumin and albumin mRNAs extracted from immunoprecipitates are enriched for the synthesis of their respective proteins by 1.8and 8.7-fold, exactly the degree of purification predicted from their relative rates of synthesis; (d) isolation of albumin synthesizing polysomes from a mixture of rat liver polysomes and hen oviduct polysomes resulted in nonspecific precipitation of only 0.4 % of the ovalbumin synthesizing polysomes when the immunoprecipitated RNA was assayed for its ability to synthesize ovalbumin in vitro. Indirect immunoprecipitation appears to be a general to the separation and isolation of polysomes and messenger RNAs coding for specific proteins.
A major approach to the investigation of differential gene expression and regulation of protein synthesis in animal cells is through the study of messenger RNA.
Lost of these mRNAs code for proteins which represent a major fraction of cell protein synthesis.
They have usually been separated from other mRNAs by fractionation on the basis of size. However, most mRNAs, especially those which are not unique in size or code for proteins representing a small percentage of cell protein synthesis, cannot be isolated free of contaminating mRNAs by size fractionation. We have developed an immunological approach based on the ability of antibody prepared against native protein to react with nascent peptide chains in polysomes.
Our previous studies have demonstrated that labeled anti-ovalbumin and antialbumin bind specifically to their respective polysomes (14, 15). Ovalbumin-synthesizing polysomes may be precipitated by sequential addition of ovalbumin and anti-ovalbumin (16) or by immunoadsorption to a cross-linked ovalbumin matrix (17). The three primary considerations which have emerged during development of these immunological methods of polysome isolation are: (a) the amount of nonspecific contamination; (b) the yield of isolated polysomes; (c) the amount of antigen and antibody required.
Direct precipitation and immunoadsorption both require large amounts of pure protein antigen which is difficult to obtain in most cases. Direct precipitation requires a large excess of antibody and produces a large precipitate illcreasing the possibility of extensive contamination by adsorbed polysomes.
Only about one-third of the available polysomes were isolated by immunoadsorption (17). This paper describes a polysome precipitation method based on indirect immunoprecipitation which overcomes these problems. Antibody against the purified native protein is reacted with polysomes and the antibody-nascent chairl-polysome complex is precipitated with an anti-antibody.
The small amount  of antigen  required  and extremely  high specificity  and yield should make t,his method   widely  applicable  to ccl1 and viral mRiYAs  coding  for the many  important,  proteins  which  constitute  less than 20% of cell  chaili-polysome complex with a second antibody, prepared against the first antibody (e.g. an anti-antibody) ; (c) sedimentation of the insoluble antibody-antibody-polysome complex through a detergent-containing discontinuous sucrose gradient to remove nonspecifically adsorbed material. The RNA is then extracted from the precipitates using sodium dodecyl sulfate and EDTA (17).
In characterizing the conditions for the second antibody reaction, we found that 1 mg of anti-antibody precipitated 30 pg of anti-Ov at the equivalence point.
Rapid, quantitative, and reproducible precipitation of anti-Ov, anti-BSA, and anti-RSA was achieved using 1 mg of anti-antibody per 15 pg of rabbit antibody and this ratio was used in all subsequent studies.
To define the reaction conditions, oviduct polysomes containing labeled nascent peptide chains were employed.
The optimum concentration of anti-ovalbumin for indirect immunoprecipitation of ovalbumin polysomes is governed by two factors; the fraction of ovalbumin polysomes precipitated and the amount of nonspecific precipitation.
To determine the amount of nonspecific precipitation, antibody against a non-oviduct protein (anti-BSA) was added to oviduct polysomes with labeled nascent chains and the radioactivity in indirect immunoprecipitates determined.
In the experiment shown in Fig. 2, the concentration of anti-Ov or anti-USA was varied and the radioactivity precipitated with each antibody determined. The optimum ratio of specific (anti-Ov) to nonspecific (anti-USA) precipitation was at about 125 pg of anti-Ov per ml. Two methods were developed to further reduce the degree of nonspecific binding.
A decrease in the amount of primary antibody used and hence the size of the precipitate produces a commensurate decrease in the amount of nonspecifically adsorbed material. 2 We therefore separated the unreacted antibody from the antibody-polysome complex by reisolation of the polysomes prior to addition of goat anti-rabbit antibody. After dialysis to remove sucrose, the minimum amount of anti-Ov (20 pg per ml) necessary for precipitation of preformed antibody-polysome complex by anti-antibody was added. antibody was then added and immunoprecipitation carried out. Although reisolation and subsequent dialysis resulted in a significant loss of polysomes, substantial improvement in the ratio of specific to nonspecific precipitation was achieved (Table  I). Due to the complexity of this procedure, it was not employed in any of the studies described.
Resuspension of the precipitate and resedimentation through the discontinuous sucrose-detergent gradient produced little loss of specific precipitate and decreased the amount of nonspecific precipitate by almost 50 y0 (Table I). This method was therefore employed in all subsequent experiments.
Speci$city of Indirect Immunoprecipitation-The specificity of indirect immunoprecipitation was demonstrated by comparison of the nascent, polypeptide chains associated with total polysomes and specific and nonspecific immunoprecipitates (Fig. 3). The specific immunoprecipitate contains essentially no radioactivity in polypeptides larger than ovalbumin. The nascent chains nonspecifically adsorbed to the anti-BSA-anti-antibody precipitate exhibit a pattern on sodium dodecyl sulfate acrylamide gel electrophoresis which parallels that of total nascent chains 0%. 3), indicating that nonspecific precipitation represents random low level adsorption of labeled polysomes.
The specificity of indirect immunoprecipitates was also investigated by examination of precipitated RNA. RNA was separated from protein and heparin in precipitates by sodium dodecyl sulfate extraction and sodium dodecyl sulfate-sucrose gradient ultracentrifugation.
No RNA was visible when anti-USA was used instead of anti-Ov (Fig. 4). In vitro synthesis demonstrated that RNA extracted from the nonspecific BSA precipitate contained less than 1 y0 of the ovalbumin synthesizing activity of RNA from anti-Ov precipitates.
The precise ovalbumin synthesizing activity of nonspecific immunoprecipitates was so low it could not be determined accurately.
A direct demonstration that indirect immunoprecipitation results in separation of ovalbumin polysomes and mRNA from total polysomes was achieved by measurement of the capacity of RNA isolated by indirect immunoprecipitation to direct ovalbumin synthesis in a heterologous cell-free protein synthesizing system. The specific activity (counts of ovalbumin synthesized per pg of RNA) of RNA extracted from purified ovalbumin  5A). Raising the ovalbumin antibody concentration permits the nearly quantitative isolation of ovalbumin synthesizing polysomes (Fig. 5B). Almost 90% of polysomal ovalbumin mRNh was recovered by indirect immunoprecipitation at the higher anti-Ov concentration (Fig. 5B). However, the amount of nonspecific precipitation is significantly increased at this concentration of rabbit anti-Ov (Fig. 3) of the increase in specific albumin synthesizing activity and that expected from its rate of synthesis provides a clear demonstration of the polysome purification achieved. (b) To isolate albumiIl-eSIlthesizirlg polysomes from a misture of hen oviduct and rat liver polysomes and measure the extent of nonspecific precipitation of ovalbumin polysomcs by assaying the immunoprecipitated RXA1 for its ability to synthesize albumin and ovalbumin in vitro.
13y immunochemical precipitation of labeled albumin from rat liver homogenates, we have determined that albumin synthesis accounts for 11.3y0 of liver protein synthesis (see "Materials and Methods"). This is in excellent agreement with the figure of 10.9% reported by Peters and Peters (20). Separation of the polysome for a protein representing 11.3 70 of cell protein synthesis from other polysomes should result in an M-fold increase in the specific synthesizing capacity of the purified RNA.
Rat liver albumin synthesizing polysomes were isolated by incubation with rabbit anti-R&~ followed by indirect immunoprecipitation with anti-antibody. The ability of RNA extracted from immunoprecipitates to synthesize albumin was assayed in the in vitro rabbit reticulocyte protein synthesizing system. The specific albumill-synthesizing activity (counts per min of albumin synthesized per pg of RNA) of the RNA in the immunoprecipitate was 8.7.fold greater than the specific activity of RNA extracted from total rat liver polysomes (Fig. 6).
The observed 8.7.fold increase in the specific aibumin-synthe- sizing activity of the precipitated RN,1 is in excellent agreement with the 8.8.fold increase predicted from the independently measured rate of albumin synthesis.
The ideal measurement of nonspecific precipitation would be to determine the amount of nonspecific mRNA isolated in the indirect immunoprecipitation of a specific polysome.
We have approached this question by measurement of the amount of ovalbumin mRTc'A precipitated during the isolation of albumin polysomes from a mixture of equal amounts of hen oviduct and rat liver polysomes. @albumin and albumin mRNAs can be assayed in, vitro simultaneously without detectable cross-reaction.* Only 0.43yl of the ovalbumin synthesizing activity was precipitated (Table 11). Since the oviduct polysomes which u-ere not precipitated retained over 9O70 of their ovalbumin synthesizing activity (Table II) and the precipitated RNA was enriched &!-fold for rat liver albumin synthesis (Table II), ovalbumin rnRIi*-l could not have been nonspecifically precipitated and escaped detection because of degradation during isolation.
Nonspecific precipitation apparently results from random low level adsorption of polysomes (Fig. 4). Therefore, the level of nonspecific adsorption of total polysomes should be similar to the 0.470 observed for ovalbumin polysomes.
The purity of the precipitated albumin mRNA can be calculated from the amount of RN.4 precipitated, and the fraction of the precipitate due to nonspecific adsorption of 0.4% of total polysomes. These results indicate that albumin mRN,4 isolated from liver polysomes is greater than 95oj, pure with respect to contamination by other mRNAs.3 3

DISCUSSION
The major goals of this study were (a) to extend our previous immunological approach to polysome isolation so as to make it applicable to a broad range of cell proteins, especially those which represent a small percentage of cell protein synthesis; (b) to define in a rigorous way the degree of nonspecific contamination of isolated polysomes; (c) to isolate rat liver albumin synthesizing polysomes and demonstrat'e the applicability of indirect immunoprecipitation to a protein which is a much lower percentage of total protein synthesis than ovalbumin.
The reaction conditions for the separation of ovalbumin and albumin polysomes from other polysomes and mRNAs were optimized by precipitation of oviduct polysomes containing leucine labeled nascent chains with specific (anti-Ov) and nonspecific (anti-BSA) antibodies. A second sedimentation of the indirect immunoprecipitates through a detergent containing discontinuous sucrose gradient greatly reduced the degree of nonspecific binding (Table I). Since this procedure is relatively simple and results in very small losses, it was used throughout these studies.
The amount of nonspecific adsorption in precipitates appears roughly proportional to the size of the immunoprecipitate, not to the amount of polysomes precipitated (Fig. 2). To minimize the size of immunoprecipitates, we separated oviduct polysomes from unreacted antibody by reisolation of the polysomes.
Although reisolation resulted in a significant loss of polysomes, it did improve the ratio of specific to nonspecific immunoprecipitate. It seems probable that polysome losses on reisolation were due primarily to the small amount of polysomes used and that losses would be minimized if reisolation was used in large scale preparative immunoprecipitations.
Although it was not employed in these studies, reisolation of the polysomes and addition of the minimum amount of antibody required to form a small precipitat,e should prove useful for isolation of polysomes synthesizing proteins which represent a very small percentage of cell protein synthesis.
Polysome reisolation also permits use of unpurified antibodies, since the large excess of unreacted nonspecific antibody is separated from the antibody-polysome complex on reisolation.
Quantitation of the degree of nonspecific binding and demonstration of the specificity of indirect immunoprecipitation were achieved by the following experiments.
(a) Reaction of oviduct polysomes containing labeled nascent chains with an&BSA results in precipitation of only 0.470 of the labeled polysomes (Table I). (b) Although oviduct polysomes possess many labeled nascent peptide chains larger than ovalbumin, indirect immunoprecipitates show essentially no nascent chains larger than ovalbumin and exhibit the pattern expected for nascent ovalbumin chains on sodium dodecyl sulfate-acrylamide gel electrophoresis (Fig. 3). (c) Indirect immunoprecipitates prepared with oviduct polysomes and anti-BSA contain no detectable RNA and have less than 170 the i n vitro ovalbumin synthetic capacity of polysomes precipitated with anti-Ov (Fig. 4). (d) The capacity of precipitated oviduct polysomes to synthesize ovalbumin mRNA is increased by 1.7.fold (Fig. 5), exactly as expected for a protein comprising 60% of protein synthesis (1.66.fold).
(e) Since ovalbumin represents 60% of oviduct protein synthesis, we extended these observations to rat liver albumin which represents only 11.3% of liver protein synthesis.
The 8.7.fold enrichment achieved is in excellent agreement with the value of 8.8. fold predicted for the purification of a polysome synthesizing a protein comprising 11.3% of cell protein synthesis.
(f) Isolation of albumin-synthesizing polysomes from a mixture of hen oviduct and rat liver polysomes, followed by i n vitro assay of the capacity of the isolated albumin mRNA to synthesize albumin and ovalbumin permits a direct demonstration of the specificity of indirect immunoprecipitation.
In two experiments, approximately 0.470 of the ovalbumin polysomes were nonspecifically precipitated (Table II).
Since nonspecific binding appears to represent uniform low level adsorption, the level of nonspecific adsorption of total oviduct or rat liver polysomes should be 0.4%. Using the 0.4% level of nonspecific adsorption which was independently determined by precipitation of in viva labeled nascent peptide chains and by in vitro translation of contaminating ovalbumin mRNA (see a and S above), the purity of the isolated ovalbumin and albumin mRNAs was calculated.3 Ovalbumin mRNA isolated by indirect immunoprecipitation is 99% pure with respect to contamination by other species of mRNA and rat liver albumin is greater than 9570 pure. While it seems unlikely that mRNA or nuclear RNA which sediments with polysomes but is not being translated in viva or in vitro contaminates our immunoprecipitates, we cannot absolutely exclude this possibility.
In contrast to direct precipitation (16) and immobilized immunoadsorbents (17), indirect immunoprecipitation does not require purified antigen in the precipitation reaction and is therefore suitable for proteins which cannot be purified in large amounts.
The amount of specific antibody required is very small and other antibodies may be used to elicit anti-antibody production.
(Goat anti-rabbit antibody prepared against anti-Ov was used to precipitate the albumin antibody-polysome complex.) The precipitate produced is relatively small and the purified antibodies used are free of ribonuclease activity.
The yield of mRNA was 60 to 70% for ovalbumin polysomes and 70 to 80% for serum albumin polysomes compared to the 30% obtained previously by immunoadsorption of ovalbumin synthesizing polysomes (17). Nearly quantitative precipitation of ovalbumin-synthesizing polysomes may be achieved by increasing the amount of anti-Ov (Fig. 5B). However, the increase in immunoprecipitate size decreases the purity of the precipitated polysomes from 99% to about 95%.
The major advantage of the indirect immunoprecipitation techniques described is that the level of nonspecific precipitation (0.4%) is extremely low relative to that obtained in earlier studies. Unenoyama and Ono (21) have published on the immunoprecipitation of albumin and catalase polysomes.
The extent of nonspecific contamination was not measured, however, and the isolated RNA was partially degraded.
Recently Schechter (22) has reported the isolation of immunoglobulin light chains by an anti-antibody method.
Although the recovery of polysomes and procedure were not described in detail, the level of nonspecific precipitation was found to be 2.0 to 2.1%. This is the same as the level of nonspecific adsorption (2.0%) on immunoadsorption of oviduct polysomes (17), and is about five times greater than that obtained by our techniques (0.499.
Since the yield of polysomes is roughly twice that obtained following previous immunoadsorption procedures, the RNA isolated by our indirect immunoprecipitation technique contains about lo-fold less contaminating mRNA.
The isolation of the mRNA coding for a single protein requires separation from other mRNAs and from rRNA.
Separation of most mRNAs from rRNA is based on the occurrence of a polyadenylate sequence at the 3' end of almost all eukaryotic mRNAs (23)(24)(25).
The polyadenylate sequence binds to cellulose (26) and nitrocellulose filters (27) and hybridizes with oligo(dT) coupled to cellulose (28, 29) or poly(U) coupled to agarose or cellulose (30). Both ovalbumin and albumin mRNAs appear to contain poly(A) sequences since they can be separated from rRNA by binding to nitrocellulose filters (12,17) and oligo(dT) cellulose (12).
Ovalbumin mRNA purified by immunoadsorption and separated from most rRNA by binding to nitrocellulose filters has been successfully used as a template for the synthesis of a complementary DNA using Rous sarcoma virus RNA-dependent DNA polymerase (31). Rat liver albumin mRNA, here separated from other liver mRNAs for the first time, will be used as template for the synthesis of a complementary DNA sequence.