Immunological studies of yeast nuclear RNA polymerases at the subunit level.

Antisera were raised against native RNA polymerases A or B, as well as against each individual subunit of RNA polymerase A from the yeast Saccharmoyces cerevisiae. The affinity spectrum of antibodies was evaluated by reacting electrophoretically separated enzyme subunits, transferred to a membrane, with 125I-labeled immunoglobulins. Alternatively, the subunit . immunoglobulin complex was revealed by 125I-labeled Protein A. Antibodies directed against native RNA polymerase A recognized the majority of the polypeptides forming the enzyme. When challenged with RNA polymerases B or C, this antibody preparation demonstrated the presence of polypeptides common to the three enzymes. A small cross-reaction was also found at the level of the large subunits of Enzyme B as well as some additional polypeptides of Enzyme C. Similar experiments with antibodies directed against native RNA polymerase B confirmed the presence of common subunits and also showed that the large polypeptides of the three enzymes share a few immunological determinants. Common subunits are AC40, ABC27, ABC23, AC19, and ABC14.5. Immunologically related sites were conserved in the large subunits of RNA polymerase A from remote yeast species. Similarly, yeast and wheat germ RNA polymerase B share immunological determinants on the large subunit as well as on a small peptide. On the other hand, there was no significant cross-reaction between yeast and mammalian Enzyme B or Escherichia coli RNA polymerase. Antibodies raised against the different polypeptide components of RNA polymerase A reacted specifically with the corresponding subunits. Inhibition studies with these subunit-specific antibodies showed that the common subunits are not always similarly exposed to antibody attack within the three enzymes. The data are discussed in terms of the structural similarity, organization and evolution of eukaryotic RNA polymerases.

Antisera were raised against native RNA polymerases A or B, as well as against each individual subunit of RNA polymerase A from the yeast Saccharomyces cerevisiae. The affinity spectrum of antibodies was evaluated by reacting electrophoretically separated enzyme subunits, transferred to a membrane, with lZ5Ilabeled immunoglobulins. Alternatively, the subunit immunoglobulin complex was revealed by '251-labeled Protein A. Antibodies directed against native RNA polymerase A recognized the majority of the polypeptides forming the enzyme. When challenged with RNA polymerases B or C, this antibody preparation demonstrated the presence of polypeptides common to the three enzymes. A small cross-reaction was also found at the level of the large subunits of Enzyme B as well as some additional polypeptides of Enzyme C. Similar experiments with antibodies directed against native RNA polymerase B confirmed the presence of common subunits and also showed that the large polypeptides of the three enzymes share a few immunological determinants. Common subunits are AC40, ABC2?, ABCz3, ACI9, and ABC14.5.
Immunologically related sites were conserved in the large subunits of RNA polymerase A from remote yeast species. Similarly, yeast and wheat germ RNA polymerase B share immunological determinants on the large subunit as well as on a small peptide. On the other hand, there was no significant cross-reaction between yeast and mammalian Enzyme B or Escherichia coli RNA polymerase.
Antibodies raised against the different polypeptide components of RNA polymerase A reacted specifically with the corresponding subunits. Inhibition studies with these subunit-specific antibodies showed that the common subunits are not always similarly exposed to antibody attack within the three enzymes. The data are discussed in terms of the structural similarity, organization and evolution of eukaryotic RNA polymerases.
Three forms of RNA polymerases, A, B, and C, are responsible for the transcription of yeast nuclear DNA (Roeder, 1976Ponta et al., 1971and Adman et al., 1972). Among eukaryotic RNA polymerases, the yeast enzymes, for practical reasons, have been the most extensively studied at the molecular level and Hager et al., 1976). These studies have revealed the extreme molecular complexity of these multisubunit enzymes which are the major component * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 4 Supported by a fellowship from the Royal Society of England.
in a system concerned with the regulation of gene expression (Huet et al., 1975;Dezelee et al., 1976;Valenzuela et al., 1976b). The majority of the subunits of the RNA polymerases have unique mass. Some polypeptides are phosphorylated in vivo (Buhler et al., 1976a;Bell et al., 1977). Structural studies have also demonstrated the presence of common subunits which belong to the pool of the small peptides. Enzymes A and B possess three subunits which are indistinguishable on the basis of molecular weight, isoelectric point, and fingerprint pattern (Buhler et al., 1976b;Sentenac et al., 1976). Enzyme C also probably shares polypeptides with enzymes A and B (Valenzuela et al., 1976a).
The existence of common subunits is a possible explanation for the previous findings that yeast RNA polymerases A and B (Hildebrandt et al., 1973;Buhler et al., 1976b) as well as the corresponding mammalian enzymes (Ingles, 1973) exhibit extensive cross-reactivity. However, it was also possible that the functionally homologous subunits were structurally related. It was suggested that the polypeptides involved in the interaction with a common subunit in the three enzymes may have retained the same subunit binding site (Buhler et al., 1976b). The fact that enzymes B and C from mammalian cells (Seifart et al., 1972;Roeder, 1976) and enzymes B and A from yeast (Huet et al., 1975) are both sensitive to a-amanitin also suggested a certain conservation of the a-amanitin binding site. Partial evidence that RNA polymerases A and B are primarily constructed of distinct gene products was again immunological. Antibodies raised against the largest subunit from RNA polymerase A do not cross-react with or inhibit RNA polymerase B (Buhler et al., 1976b). This immunological approach needed to be extended to all the individual subunits in order to present a complete picture of the structural relationship of the three forms of RNA polymerase. For this purpose, we used a protein blotting technique to allow antibodies to react with the subunits separated by SDS'-polyacrylamide gel electrophoresis. The results establish that the different subunits of yeast RNA polymerase A are unique proteins and identify the polypeptides in the three enzymes which are immunologically related. These immunological studies have been extended to RNA polymerases from other organisms.

RNA Polymerases and Subunits-Yeast
RNA polymerases A and B were purified from Saccharomyces cereuisiae as previously described (Huet et al., 1975;Dezelee et al., 1976). RNA polymerase C was purified according to Hager et al. (1977). Escherichia coli RNA polymerase was a gift of A. Ruet (this laboratory). Wheat germ and calf thymus RNA polymerase B were donated by C. Kedinger (Strasbourg University).
Subunits of yeast RNA polymerase A were isolated on a preparative scale by SDS-polyacrylamide gel electrophoresis on a 1.0-mm ' The abbreviation used is: SDS, sodium dodecyl sulfate. 9949 thick, 11.5% acrylamide gel, using 2 mg of RNA polymerase per run. The different polypeptides were visualized after electrophoresis by immersing the gel into a KC1 solution (0.1 M) in the cold for 30 min. Polyacrylamide strips containing the subunits were cut out and crushed with a glass homogenizer. A portion of each mixture was subjected to an analytical SDS-polyacrylamide gel electrophoresis to check the purity of the subunits before each injection series. A total of IO mg of RNA polymerase was needed to obtain the antibodies.
Antisera-Antibodies against native RNA polymerases A and B were raised in rabbits by injecting 250 pg of protein, emulsified with complete Freund's adjuvant (Difco), intracutaneously at several sites in the back of the rabbit. Booster injections were performed in the same way, with incomplete adjuvant (Difco), after 21 days and subsequently every 10 days. The animals were bled 12 days after the sixth and seventh injections.
Antibodies directed against the isolated subunits of yeast RNA polymerase A were obtained similarly, by injecting the immunogen under the form of a suspension of crushed polyacrylamide with 1 volume of Freund's complete adjuvant. Booster injections were done with incomplete adjuvant after 21, 36, 51, and 66 days. The protein mass injected, which depended, of course, on the molecular weight of the subunits, corresponded to 2 mg of RNA polymerase. Rabbits were bled 15 days after the fifth injection.
Purification and iodination of immunoglobulins (IgG) with were carried out according to Broome and Gilbert (1978). Protein A from Staphylococcus aureus, a gift from J. M. Dubert (University of Paris VII), was labeled with 1251 by the same procedure to a specific activity of about 4 X 10' cpm/pg.
Protein-blotting Procedures-RNA polymerases (150 pg) were frst subjected to electrophoresis in a 11.5% polyacrylamide slab gel (10 X 9 X 0.06 cm) with sodium dodecyl sulfate, according to Laemmli (1970). The subunits were then transferred to a membrane by direct blotting or electrophoresis. Direct blotting was carried out essentially as described by Southern (1975) for blotting DNA fragments. The proteins were covalently linked to a cellulose acetate membrane (SM 11106, Sartorius) activated with cyanogen bromide just before the transfer. The blotting buffer (PBS) was 0.8% NaCl, 10 m~ sodium phosphate, pH 7.5. The flow rate through the acrylamide gel was 0.2 ml/h/cm2. After 20 h of blotting, the membrane was removed, and washed for 4 h in PBS containing glycine (15 g/liter) at 4"C, then rinsed with and stored in PBS + 0.5% rabbit serum and 0.1% serum albumin until used for immunological detection. The electrophoretic transfer of RNA polymerase subunits was carried out as described by Towbin et al. (1979). Electrophoresis of RNA polymerase A under nondenaturing conditions was performed as previously described (Huet et al. 1975). The enzyme was transferred to activated cellulose acetate membrane by direct blotting as for the subunits.
Visualization ofAntibody Binding-After transfer of the proteins, the membrane was soaked in 3% bovine serum albumin in PBS for 1 h at 40"C, rinsed three times with a Triton/SDS buffer (Dimitriadis, 1979) (PBS with 0.5% bovine serum albumin, 0.2% Triton X-100, and 0.2% SDS) at room temperature. The membrane was then incubated with the '251-labeled IgG in the same buffer (10' c p m / d ; 0.15 ml/cm' of filter) into a sealed plastic bag with gentle rocking at 22°C for 17 h. The membrane was washed four times with 50 ml of the above Triton/SDS buffer for a total of 45 min, blotted on filter paper, thoroughly air dried, and subjected to autoradiography on x-ray film (3M; type XM) with intensifying screen (3M; type a-16 screen) for 4 h or overnight at room temperature. For detection with Protein A, the blot was incubated with unlabeled IgG diluted to 30 pg/ml with the Triton/SDS buffer and washed as indicated above, then incubated with "'1-labeled Protein A (5.5 X lo5 cpm/ml; 0.15 ml/cm') in Triton/ SDS buffer for 1 h at room temperature. The membrane was washed with 50 ml of Triton/SDS buffer four times, dried, and subjected to autoradiography as above.
Effect of Specific Antibodies on Enzyme Activity-Purified IgG directed against isolated subunits of RNA polymerase A were preincubated at varying concentrations with 1.5 pg of RNA polymerase A, B, or C, in 50 pl of a solution containing 50 mM Tris-HC1 (pH 7.4), 2 mM NaCI, and 0.1 m~ EDTA, for 1 h at 37°C. The final concentration of IgG was kept constant by complementing with control y-globulins from a normal rabbit. Then, the residual enzyme activity was estimated by a further incubation of 20 min at 30°C, after addition of 50 pl of the following transcription mixtures. For RNA polymerase A, it contained 150 mM Tris-HC1 (pH 8). 2 mM dithiothreitol, 10 mM MgC12, 10 pg of denatured calf thymus DNA, 2 mM each of ATP, GTP, and CTP, and 1 mM [a-'"PP]UTP (20 cprn/pmol). For RNA polymerase B, it contained 50 mM Tris-HC1 (pH a), 10 m~ dithiothreitol, 10 mM of Yeast RNA Polymerases MnCL, 100 m~ ammonium sulfate, 10 pg of denatured calf thymus DNA, 1 mM concentration each of ATP, GTP, and CTP, and 0.2 mM [a-"PIUTP (12 cpm/pmol). In this case, incubation was for 10 min at 30°C. For RNA polymerase C, it contained 50 mM Tris-HC1 (pH 8), 2 mM dithiothreitol, 2 m~ MnC12, 10 mM MgC12, 100 mM ammonium sulfate, 10 pg of denatured calf thymus DNA, 2 mM each of ATP, CTP, and GTP, and 1 m~ [a-"PIUTP (25 cpm/pmol). After incubation, acid-precipitable radioactivity was collected on a membrane filter (HAWP 025, Millipore) and measured by liquid scintillation.

RESULTS
Affinity Spectrum of Antibodies-Antibodies raised against yeast RNA polymerase A cross-react with and inhibit RNA polymerase B (Hildebrandt et al., 1973). To interpret this observation, the affinity spectrum of the antibody preparation had to be determined at the subunit level. A method was developed to react the '"I-labeled antibodies with the individual subunits of R N A polymerase. The proteins were fist separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, then transferred to an activated cellulose acetate membrane essentially as described by Southern (1975) for transfer of D N A fragments. The proteins covalently attached t o the membrane were subsequently incubated with the I2'Ilabeled immunoglobulins. Because yeast RNA polymerases are made of polypeptides of largely different mass, from 190,000 to 12,000 daltons, one could expect a preferential transfer of the small proteins. The efficiency of the transfer method was therefore quantified using a mixture of I4C-labeled proteins from yeast separated as a function of molecular weight by SDS-polyacrylamide gel electrophoresis. Basic and acidic proteins were analyzed independently. The transfer of proteins to the activated membrane was estimated by scanning the autoradiogram of the membrane and of the gel before and after blotting.
In the case of acidic proteins, the 14Cprotein pattern after transfer was similar to the original one and the extent of transfer was in the range of 50%. A more erratic transfer occurred from the pool of basic proteins. Although most of the proteins were detectable, the extent of transfer varied from protein to protein. However, this did not depend simply on the molecular size of the polypeptides. In fact, the method can be applied to very large proteins like RNA polymerase A* (Mr = 550,000) (Huet et al. 1975) after electrophoresis under nondenaturing conditions (Fig. 1). The method is sensitive enough to tolerate, in this case, a low transfer efficiency.
The affinity spectrum of the antibodies directed to native RNA polymerase A was determined after blotting the subunits on a membrane (Fig. 1). The antibody preparation recognized the majority of the subunits. The antibodies bound most efficiently subunits A1m and &. A clear binding response was obtained on Aw,, A 4 9 , AM, AZ7, AZ9, Ale, and although the signal intensity varied from one experiment to another, especially on using the same antibody preparation. One polypeptide, A43, was poorly and only occasionally detected.
Three different antibody preparations gave grossly a similar binding pattern mainly with variations in binding intensity on different subunits. The enzyme can be boiled in 1% SDS (w/ v) or concentrated by precipitation with 10% trichloroacetic acid (w/v) and still retain the same pattern of antibody recognition (Fig. 1). The low and variable binding response of subunit A2, reflected a poor transfer to the membrane as evidenced by staining the gel for proteins after blotting.
While this protein blotting method was developed, two alternative procedures were described (Renart et al., 1979;Towbin et al., 1979). We were interested in using the electrophoretic technique of Towbin et al. (1979) in comparison with our direct blotting procedure to determine whether the method of transfer influenced the apparent affinity spectrum of antibodies. The isoelectric point of RNA polymerase A subunits range from 9.2 to 4.5; thus, a nitrocellulose membrane was placed on both sides of the gel since the direction of migration of basic proteins complexed with SDS could not be predicted. Unexpectedly, almost the same subunit pattern was detected on both sides. Only small differences were noted between the electrophoretic and direct transfer techniques; subunit was more reproducibly transferred by electrophoresis, whereas Lo was always more intensively labeled after direct blotting. In both cases, An; and &:%were poorly detected. We found that a more uniform labeling of the subunit-immunoglobulin complexes could be achieved by using lZ5Ilabeled Protein A from S. aureus (Renart et al. 1979). In most cases, Azi and gave a good binding signal with Protein A which showed the presence of bound IgG onto those two subunits (see Lane 6 in Fig. 1). The different response to the two labeled probes could be attributed in part to a variable transfer but it also suggested that iodination of IgG may affect their binding efficiency.
Since the majority of the subunits were recognized by antibodies, it was possible to explore the extent of crossreaction of RNA polymerases from different origins, using a combination of the two blotting techniques.
Immunological Relationship of Different RNA Polymerases-RNA polymerase A from the yeast Schizosaccharomyces pombe, Candida tropicalis, and Endomycopsis fibuligera have a molecular structure which differ from that of S. cerecisiae in the number and molecular weight of their subunits . All are constituted of two large polypeptides with a collection of smaller chains. As shown in wheat germ, were separated by SDS-gel electrophoresis, transferred electrophoretically to a nitrocellulose sheet, and allowed to react with ""I-labeled IgG directed against S. cereuisiae RNA polymerase B.
The autoradiogram shows the subunits which have bound the labeled antibodies.
tropicalis and E. fibuligera enzymes and to a lesser extent with the second largest one. A faint signal was also detected, in another experiment, on two smaller polypeptides (the common subunits A2:r and AI9). On the other hand, there was no detectable cross-reaction with RNA polymerases from S. pombe (enzyme A), E. coli (holoenzyme), or calf thymus (enzyme B). The immunological relationship of RNA polymerases B from S. cereuisiae, calf thymus, and wheat germ was similarly investigated at the subunit level. Using ""I-labeled antibodies directed against native enzyme B from yeast, a clear cross-reaction signal was found on the largest subunit of wheat germ enzyme B, as well as on the smallest of the polypeptides (Fig. 2). Only a faint binding response was detected on the large subunits of the mammalian enz-me. There was no significant signal with E. coli RNA polymerase (not shown).
Immunological Relationship between Yeast RNA Polymerases A, B, and C-One of the purposes of this study was to examine the immunological relationship between the three forms of yeast nuclear RNA polymerases. Their subunits were separated on an SDS-polyacrylamide slab gel and challenged with antibodies raised against RNA polymerases A or B. The bound immunoglobulins were then revealed with 1251-labeled Protein A (Fig. 3). Antibodies against RNA polymerase A bound subunits B2;, and B14.5. The intensity of the binding signal on these small subunits was the same for enzymes A and B. In addition, a clear binding response was distinguished on the two large subunits, especially on Bncl (or i t s proteolyzed form, B d . However, in this case, antibody binding was much less strong than on the large subunits of enzyme A, suggesting that the cross-reactivity was limited to one or a few immunological determinants. Reciprocally, antibodies against RNA polymerase B which recognized subunits B??,,, BIN,, BI~dlr B44.5r Brrz, B2:. B2:l, BM, B14.', and also bound subunits At;, AZI, and AI.,.R from enzyme A with the same efficiency as for Bt;, Bnn, and BI4.'. In addition, a small cross-reaction was distinguished on Ala, and B. and C (10 pg) were separated by electrophoresis on polyacrylamide gel with SIX. transferred by electrophoresis on a nitrocellulose membrane, and incubated with antibodies directed against R N A polymerase A (anti-A) or against R N A polymerase B (anti-B). The subunit. IgC complexes were then revealed by incubation with ",.'I-labeled Protein A and the membrane was subjected to autoradiography as described under "Experimental Procedures." The enzymes are identified by the letters A. E, or C and the subunits by their molecular weight X 10 .I.
In another experiment, RNA polymerase C subunits' were probed with antibodies against enzyme A or B (Fig. 3). Anti-RNA polymerase A antibodies revealed subunits CJll, c?;, and CZ:~ with the same efficiency as AJll, A?;, and hence demonstrating the immunological relationship of these subunits. Als and A14.9 were poorly detected in this experiment but another antibody preparation cross-reacted with these two putative common subunits in enzyme C (results not shown). Interestingly, another polypeptide of our enzyme C preparation, Ca:l, bound antibodies. This polypeptide was designated as a component of RNA polymerase C by Hager et al. (1977). There was also a weak binding signal on other polypeptides, CIZ", CJ9, and c : l 4 but these results need further investigation.
Anti-RNA polymerase B revealed subunits Cr; and CZrl. A weak cross-reaction signal also revealed C5,1, suggesting an immunological relationship with one of the B subunits. These experiments were repeated using the direct transfer technique or the electrophoresis, with similar results.
Studies using Antibodies Directed against Individual Subunits-To probe in more detail the structure of yeast RNA polymerases, antisera were raised against each individual subunit of RNA polymerase A, separated by gel electrophoresis. The reactivity of the immunoglobulins towards the polypeptide components of enzymes A and B was evaluated by the electrophoretic transfer technique (Fig. 4). Both a short (4 h) and long (24 h) exposure time for autoradiography was used to estimate the extent of specificity of the antibodybinding response. The majority of the SDS-denatured subunits of enzyme A were specifically recognized by their respective antiserum. This was the case for Alml, AI.13, Ado, A:I~B, A~T, Az:~, and Ats. Anti-Alal IgC also reacted with a protein band, migrating in between Alw and which was originally thought to be a contaminant. This proteolytic degradation product of AIw was only detected on the membrane placed on the cathodic side. Limited proteolysis of A49 may ? We use our nomenclature throughout for the purpose of homogeneity (Sentenac et af., 1976;Buhler et af., 1976b). The subunits' molecular weights reported by Hager et al.. (1977)  FIG. 4. Specificity of antisera directed against the isolated subunits of RNA polymerase A. IgC directed against subunits of R N A polymerase A were obtained, purified, and labeled with '''1 as described under "Experimental Procedures." Purified yeast R N A polymerase A (150 pg) was denatured with SDS, the subunits were separated in a SIIS-polyacrylamide slab gel by electrophoresis and transferred to a nitrocellulose membrane. The membrane was cut and individual strips were exposed to "'I-labeled IgC directed against the different subunits. The autoradiogram shows, from feft to right, the subunits revealed by antibodies against subunits A,w. All;. Aw, AI.1. also explain the weak bands of smaller molecular weight revealed upon long autoradiography. Anti-&:l IgC reacted poorly with A,,:,. The immunoglobulins gave instead a strong signal with A,!, and reacted also with A40 and AM. Since antibodies against these polypeptides did not reciprocally recognize A4:,, it is likely that the preparation injected into the rabbit was contaminated with these very antigenic proteins and possibly by a proteolytic product from There is no straightforward explanation for the weak signals occasionally detected on bands of higher molecular weight than the antigen. Hence, analysis of anti-(Al,.:, + AI,) and anti-Al2.L. IgG was repeatedly hampered by a fuzzy background. In addition to the corresponding subunit, undefined bands of higher molecular weight were detected. After overexposure of the film, anti-A19 gave also a weak but discrete signal on three additional protein bands larger than Aw. This may be the case where the existence of antibodies specific to SDS or SDStreated protein creates spurious background and nonspecific binding of antibodies (Lompre et al., 1979). This is more likely to be a problem for the small subunits because a large band of SDS-polyacrylamide gel had to be injected into the rabbit.
The subunit-directed antibodies were used to probe the activity of native RNA polymerases A, B, and C (Fig. 5 ) . The effect of the specific antibodies on RNA pol-ymerase A activity is shown in Fig. 5A. The antibodies can be separated into two main groups, with different inhibitory properties. The most inhibitory group consists of anti-Al%l, -AI:I~, -AJo, -A~:I, -AM, and -(Al4.5 + A14) immunoglobulins. Complete inhibition of transcription could be achieved upon preincubation of enzyme with excess antibodies. The mechanism of this inhibition will be the subject of another communication. The second group of antibodies is represented by anti-&!,, -Aw,, -AJ:!. and -A?; IgC. A 10-fold higher concentration of immunoglobulins was required to achieve 50% inhibition of RNA synthesis, as compared to the first group. Complete inhibition could not be obtained. This was partly expected since the removal of A49 and A:,,:, subunits only reduces by a factor of 2 the transcription of denatured DNA (Huet et al., 1975). The weak inhibitory properties of anti-APi was more unexpected. The effect of the same antibodies on the activity of RNA polymerases B and C is shown in Fig. 5. RNA

A N T I -R N A P O L Y M E R A S E A SUBUNITS IgG ( P g )
FIG. 5. Effect of antibodies against isolated subunits of RNA polymerase A on the activity of RNA polymerases A, B, and C. Purified IgG directed against individual subunits of RNA polymerase A were preincubated with RNA polymerases A, B, or C and the residual enzyme activity was measured as described under "Experimental Procedures." A , RNA polymerase A; B, RNA polymerase B; other antibodies were inactive, including those directed against the common subunits A27 and A14.5. RNA polymerase C was inhibited by anti-ho, anti-Azs, and anti-A19 immunoglobulins, practically to the same extent as RNA polymerase A. RNA polymerase C was even more sensitive to anti-Alg IgG than enzyme A. Again, the other antibodies were inactive.

DISCUSSION
Two important general conclusions can be drawn from these immunological studies. The fist one concerns the uniqueness of the polypeptide components of yeast RNA polymerase. Our preliminary fingerprint data suggested that most of the subunits of RNA polymerase A were distinct proteins and did not derive from limited proteolysis of the larger molecular weight polypeptides (Buhler et al., 1976b). This is now strongly confirmed by the specific reactivity of antibodies against isolated subunits. It is clear that AIw, A135, L9, A40r A:j4.5, Ali, A23, and AI9 are unique proteins which carry unique immunological determinants. The data do not yet allow an extension of this conclusion with certainty to the smallest subunits A14 Tr and Ala,, but it is likely that these polypeptides are also distinct gene products. If one extrapolates these results to RNA polymerases B and C, the transcription machinery in yeast appears to be one of the most complex enzymatic systems of the cell with some 25 distinct proteins.
The second conclusion concerns the structural relationship of RNA polymerases. It was known that the multiple forms of RNA polymerases are immunologically related (Hildebrandt et al., 1973). However, it was not clear whether this overall cross-reaction was due to the presence of common subunits or to the conservation of large common structural domains, or both. We have previously shown that antibodies against the largest subunit of RNA polymerase A do not noticeably crossreact with nor inhibit RNA polymerase B (Buhler et al., 1976b). This study has now been extended to all the subunits of RNA polymerase A and the cross-reaction has been studied at the subunit level with a highly sensitive technique. Two different types of antibodies were used as a probe. Antibodies against native RNA polymerase are expected to recognize the immunological determinants accessible on the enzymes under their native configuration. Antibodies against SDS-denatured C, RNA polymerase C. The different antibodies are identified by the name of the subunits used as antigen: Alw, etc. The results are given as per cent of control UMP incorporation, in the presence of control y-globulins, which was 1.5, 1.2, and 0.9 nmol by RNA polymerases A, B, and C, respectively. subunits will be directed against all the immunoreactive sites of the denatured polypeptide. The conclusion which emerges clearly from the data is that the immunological relationship between the three forms of yeast RNA polymerase is mostly accounted for by the existence of common subunits. However, the large polypeptides of the three enzymes share a few immunological determinants.
Let us examine first the case of the common subunits. On the basis of identical molecular weight, the putative common subunits are AC40, ABCZi, ABCza, ACIU, and ABCM (Buhler et al., 1976b;Valenzuela et al., 1976a). Now, on the basis of antibody binding, the immunological relationship of the following subunits is firmly established: AC4Cl, ABCZ;, ABCe.l, AC19, and ABC14.6. Inhibition data with antibodies directed against RNA polymerase A subunits confirm the structural relationship of AC4", ABCz:%, and ACI9. It also suggests that these polypeptides are similarly exposed to antibodies within the three enzymes. Hager et al. (1976) and Valenzuela et al., (1976a) have reported the co-migration, by two-dimensional gel electrophoresis, of subunits Cm and Am, as well as of and A19, while CZ3 was found to be less acidic than AB2:j. The charge difference between ABza and CZn probably stems from a different degree of phosphorylation or another covalent modification of the polypeptide. The same authors also could not detect CZi on their two-dimensional subunit map. Therefore, the only indication that Czi is identical to Ay; and Be; was their strikingly similar migration during SDS-polyacrylamide gel electrophoresis. The good binding response of Cs; to the antibodies directed against enzymes A or B is now a convincing proof of identity. A covalent modification of Cr7 may explain the abnormal electrophoretic migration in urea. There remains the case of anti-A2; and anti-(A14n + Ai4) antibodies which only inhibit RNA polymerase A, although there is convincing evidence that APi and A14 are common to the three forms of enzyme (Buhler et al., 1976b;and this work). The lack of inhibition probably reflects a different setting of these subunits within enzymes A, B, and C, thereby being less accessible to antibody attack.
There were cases of antibodies cross-reacting with polypeptides other than the previously recognized common subunits. Hence, anti-polymerase A antibodies recognized immunolog-ical determinants in subunit CS3. It should be noted that the subunit composition of RNA polymerase C is still open to question. Some polypeptides which copurify with the enzyme, like C40.s (Valenzuela et al., 1976b) or Cs3 and Csi (Hager et al., 1977) were later thought to be contaminants (Valenzuela et al., 1976a;and Bell et al., 1977). In light of the above results, it is suggested that Cs belongs to the enzyme structure. It is not clear yet to which subunit of enzyme A C53 is related.
The small but discrete cross-reaction at the level of the large subunits of enzymes A and B and to a lesser extent of enzyme C deserves some comments. The antibody-binding response was at least 1 order of magnitude lower than on the large subunits of the native enzyme used as antigen. Therefore, very few immunological determinants are conserved. The two large subunits, in all likelihood perform, in the three enzymes, the same basic steps of RNA synthesis and interact with the common polypeptides. Nevertheless, widely divergent polypeptides were probably evolved from a common ancestral set of proteins to meet the different functional and regulatory requirements of the enzymes which are located in different cell compartments and probably interact with different chromatin components. The lack of inhibition of RNA polymerase B or C by anti-AlW IgG or anti-A1:3a IgG, which confirms our preliminary observation (Buhler et al., 1976b3, is probably due to the low concentration of the antibodies recognizing the common sites. Further studies are in progress using antibodies against isolated subunits of enzymes A and B to explore in more detail the immunological relationships of the large subunits of the three forms of RNA polymerase. The conservation of immunologically related sites in the large subunits of enzyme A from remote yeast species suggests that a high degree of functional constraint was exerted on these subunits, especially on AlW which appears the most immunologically conserved. Similarly, yeast and wheat germ RNA polymerase B share immunological determinants on their largest subunit as well as on a very small peptide. This is in keeping with the Observation that antibodies directed against Drosophila RNA polymerase B inhibit yeast and calf thymus enzyme B (Greenleaf et al., 1976). On the other hand, we found no significant cross-reaction between yeast enzymes B and E. coli or mammalian RNA polymerase B. It is interesting in this respect to note that yeast and wheat germ RNA polymerase B are not competent to transcribe accurately a mammalian gene when added to a crude mammalian reconstituted system (Wed et al., 1979).