AUTOIMMUNITY TO RNA POLYMERASE II IS FOCUSED AT THE CARBOXYL TERMINAL DOMAIN OF THE LARGE SUBUNIT

Abstract

Objective

Previous studies have demonstrated antibodies to the large (220 kd) polypeptide subunit of RNA polymerase II (Pol II) in sera from certain patients with scleroderma. In the present study, we sought to identify the autoantigenic region on this polypeptide.

Methods

A recombinant fusion protein, corresponding to the 52-heptapeptide repeat found in the carboxyl terminal domain (CTD) of the large Pol II subunit, was used to identify 15 patient sera that contained autoantibodies. Synthetic peptides CTD7 (representing a single heptapeptide) and CTD18 (representing 2½ heptapeptide repeats) were used in a competitive inhibition assay to define the specificity of these sera and the importance of the CTD as an autoantigen.

Results

All 15 sera immunoprecipitated the Pol II subunit from radiolabeled cell extracts, and 11 of them bound the CTD fusion protein in immunoblots. Immunoprecipitation of Pol II was completely inhibited by CTD18 in 5 sera and partially inhibited in 4 additional sera.

Conclusion

These results indicate that the CTD heptapeptide repeat is a focal point for autoimmune responses in scleroderma. It is likely that the repetitive sequence and high content of charged residues of this structure contribute to its role as an autoantigen.

Autoantibodies to RNA polymerases (Pol I, II, III) of mammalian cells appear in certain patients with scleroderma (1–5). We and others have been particularly interested in autoantibodies to Pol II (5,6). This holoenzyme transcribes premessenger RNA. It consists of 2 large polypeptides of 220 kd and 145 kd and at least 6 smaller subunits, some of which may be shared with RNA polymerase I (which transcribes ribosomal RNA) and RNA polymerase III (which transcribes various small RNAs) (7–9). In our earlier analysis, 13 of 278 patients with scleroderma were found to have autoantibodies to Pol II. Among these 13 sera, 9 bound the 220-kd polypeptide and 11 bound the 145-kd polypeptide in immunoblots (5). Only 1 serum immunoprecipitated Pol II without recognizing either of these polypeptides.

In the course of that study, we noted that a spontaneously occurring 180-kd breakdown product of the largest Pol II subunit was not recognized by patient sera. This fragment is known to be derived from the 220-kd polypeptide through loss of its carboxyl terminal domain (CTD), which consists of a Tyr-Ser-Pro-Thr-Ser-Pro-Ser heptapeptide that is repeated 52 times in the Pol II of mammalian cells (8,10–13). This observation suggested that the CTD might represent a principal epitope for autoimmune responses to Pol II. In the present study, we investigated this possibility in some detail, using the original 13 patient sera plus 2 additional sera identified subsequently.

PATIENTS AND METHODS

Patient sera

Sera were obtained from 13 adult patients known to have antibodies to Pol II (5) and 2 additional adult patients in whom these antibodies were identified by immunoprecipitation assay. These patients were followed up in clinics at the University of Pittsburgh, Keio University, and Yale University. All 15 (13 women, 2 men; 13 of white and 2 of Japanese ancestry) met the American College of Rheumatology (formerly, the American Rheumatism Association) preliminary criteria for scleroderma (14).

A murine monoclonal antibody (MAb) specific for the CTD of the largest subunit of Pol II (8wgl6; a kind gift from Dr. Nancy E. Thompson, McArdle Laboratory for Cancer Research, University of Wisconsin, Madison) was used as a positive control (15,16). Anti-Sm Mab Y12 and sera from patients with systemic lupus erythematosus (SLE), patients with polymyositis/dermatomyositis, and normal blood donors were used as negative controls (17).

Radioimmunoprecipitation

Antibodies to RNA Pol II were detected in radioimmunoprecipitation assays performed as described previously (18,19). Briefly, HeLa cells were radiolabeled for 14 hours with ³⁵S-methionine (10 μCi/ml cells; ICN Biomedical, Costa Mesa, CA). Cell extracts were prepared using sonication. Immunoprecipitation was performed using 10 μl of patient sera per 2 mg of protein A–Sepharose beads (Pharmacia, Piscataway, NJ) in 500 μl of immunoprecipitation buffer (IPP; 10 mM Tris HC1, pH 8.0, 500 mM NaCl, 0.1% Nonidet P40). Each 2 mg of antibody-coated beads was mixed separately with 100 μl of ³⁵S-methionine–labeled cell extract (corresponding to 8 × 10⁶ cells) and rotated at 4°C for 2 hours. After 4 washes with IPP buffer, the beads were resuspended in sodium dodecyl sulfate (SDS) sample buffer and proteins were fractionated in SDS–7.5% polyacrylamide gels (20), which were analyzed by autoradiography.

Complementary DNA encoding CTD

Complementary DNA (cDNA) clones encoding the Pol II large subunit CTD were obtained using the polymerase chain reaction (PCR). Primers were based on the mouse sequence of CTD (21). The forward primer was ATG GAT CCT CTA GAT CAC CAG GTG CTA TGT CT; the reverse primer was ATG GAT CCG GGC CCT CTA GAG TTC TCC TCA TCG CTG TCA TC. These primers included terminal Bam HI sites to facilitate subcloning into pGEX-2T (Pharmacia). Amplification was performed on mouse liver genomic DNA in a thermal cycler (Perkin-Elmer/Cetus, Norwalk, CT) programmed as follows: denaturation for 1.5 minutes at 94°C, annealing for 2 minutes at 55°C, and extension for 3 minutes at 72°C. After 35 cycles, reaction mixtures were incubated for 10 minutes at 72°C to ensure completion of template synthesis. The PCR product was isolated in an agarose gel and subcloned into the pGEX-2T vector. The recombinant vector included the expected segments corresponding to pGEX-2T (4,948 basepairs) and CTD coding sequences (1,164 bp). Sequence analysis using the dideoxy chain-termination method (22) confirmed that the PCR product was identical to the previously described CTD of the mouse Pol II large subunit (21). It should be noted that the mouse and human CTD sequences differ only at 2 amino acid residues (13).

CTD fusion protein

The cDNA encoding the complete mouse CTD was expressed as a glutathione-S-transferase (GST) fusion protein referred to as CTD-GST. This protein was isolated using a glutathione affinity column (23). It was further purified through electroelution from an SDS–7% polyacrylamide gel. The resulting fusion protein migrates as a 98-kd polypeptide in SDS–polyacrylamide gel electrophoresis. Immunoblots were performed using a modification of the procedure described by Towbin et al (24).

Synthetic peptides

Synthetic peptides CTD7 (NH₂-Tyr-Ser-Pro-Thr-Ser-Pro-Ser-amide) and CTD18 (NH₂-[Tyr-Ser-Pro-Thr-Ser-Pro-Ser]₂-Tyr-Ser-Pro-Thr-amide) based on the CTD consensus sequence were synthesized with the solid-phase method using protocols recommended by the manufacturer (25). These peptides were tested over a range of concentrations to identify the optimal quantity for inhibiting MAb 8wgl6 and prototype patient serum Wa in immunoprecipitation. Inhibition was not observed with CTD7. For CTD18, optimal conditions were as follows. Protein A–Sepharose beads (2 mg) were coated with IgG from 5 μl of patient sera and were incubated with 20 μg of peptide in 400 μl of IPP buffer for 2 hours. After 5 washes with IPP buffer, the beads were combined with 100 μl of standard HeLa cell extract and 400 μl of IPP buffer, and immunoprecipitation was carried out as described above.

RESULTS

Initially, we assessed the ability of the 15 anti–Pol II positive sera to recognize the CTD-GST fusion protein in immunoblots (Figure 1). The control anti-CTD MAb 8wgl6 and prototype patient serum Wa recognized this fusion protein readily, whereas control anti-Sm MAb Y12 did not (Figure 1A). Among the 15 patient sera, 11 bound the fusion protein (as shown by 7 representative examples in Figure 1B). No binding occurred with control sera from other patients or normal donors. None of these sera bound to GST alone (results not shown). Therefore, these 11 sera were judged to contain antibodies to CTD-specific epitopes.

Figure 1.

Detection of anti–carboxyl terminal domain (anti-CTD) antibodies in immunoblots. A, Studies with known positive and negative controls. Lane 1, Amido black–stained substrate, demonstrating the 98-kDa CTD–glutathione-S-transferase fusion protein. Lane 2, Immunoblot with monoclonal antibody (MAb) 8wgl6. Lane 3, Immunoblot with prototype serum Wa. Lane 4, Immunoblot with negative control anti-Sm MAb Y12. B, Representative examples of patient sera. Lanes 1–7, Patient sera that immunoprecipitated Pol II. Lanes 8–12, Control patient sera. Lanes 13 and 14, Normal control sera.

To assess the importance of the CTD as an autoantigen, we tested the ability of synthetic peptides CTD7 (representing a single heptapeptide) and CTD18 (representing 2½ heptapeptide repeats) to inhibit immunoprecipitation of Pol II. CTD7 produced no inhibition of immunoprecipitation (results not shown). Therefore, subsequent studies were performed with CTD 18. As seen in Figure 2, CTD18 completely inhibited MAb 8wgl6 (lanes 1 and 2) and sera from patients Wa (lanes 3 and 4) and Ku (lanes 7 and 8) and partially inhibited serum from patient Kn (lanes 5 and 6). This peptide did not interfere with immunoprecipitation of the Ku (p70/p80) and Sm autoantigens (as shown in lanes 7–10), confirming its specific inhibitory effect for antibodies directed against CTD-related epitopes. Since CTD7 does not interfere with patient autoantibodies, we assume that the specific epitope spans the junction of a heptapeptide repeat. Additional peptide constructs will be needed to identify a more limited CTD sequence that inhibits these autoantibodies.

Figure 2.

Inhibition of Pol II immunoprecipitation with synthetic carboxyl terminal domain 18 (CTD₁₈) peptide. Immunoprecipitation of ³⁵S-methionine–labeled HeLa cell extracts was carried out in the presence and absence of the CTD_I8 peptide. Known Pol I, Pol II, and Pol III components are identified according to their apparent molecular weight. Asterisks indicate Pol II polypeptides. The ∼200-kd doublet shown in lanes 9 and 10 is a known component of the U5 small nuclear RNP particle recognized by anti-Sm antibodies. mAb = monoclonal antibody

It should be noted that serum Wa immunoprecipitated a prominent band at ∼190 kd and another band at 126 kd, which are known components of RNA polymerase I. Among the 15 anti–Pol II positive sera examined in this study, 13 contained antibodies to this polymerase, as judged by immunoprecipitation of these proteins. In addition, all 15 sera contained antibodies to RNA polymerase III, as judged by immunoprecipitation of proteins of 155 kd and 138 kd, which are known to be components of this enzyme. It is clear from the results shown in lanes 3–4 and 5–6 of Figure 2 that CTD 18 does not interfere with the immunoprecipitation of Pol I and Pol III.

The ability of CTD18 to act as a competitive inhibitor of immunoprecipitation for all 15 sera studied is summarized in Table 1. Inhibition with the positive control MAb 8wgl6 was complete with 5 patient sera and partial with an additional 4. It should be noted that the 4 patient sera that immunoprecipitated Pol II but did not recognize the CTD-GST fusion protein (sera 10, 11, 14, and 15) exhibited no reduction in their ability to immunoprecipitate the 220-kd band in the presence of CTD18. Therefore, these sera must contain antibodies to other regions of the large subunit or to other polypeptide components of Pol II. We conclude that most sera were responding to the CTD, and in about one-third, autoimmune responses to Pol II were confined exclusively to this molecular region.

Table 1.

Autoantibody specificities for RNA polymerase II^*

		35S-methionine IPP
Sample	Immunoblot/220 kd	HeLa extract alone	HeLa extract with CTD18	Inhibition
MAb 8wg16	++	+	−	Complete
Serum 1	++	++	−	Complete
Serum 2	++	++	−	Complete
Serum 3	++	++	−	Complete
Serum 4	++	++	±	Partial
Serum 5	++	++	±	Partial
Serum 6	+	++	+	Partial
Serum 7	+	+	−	Complete
Serum 8	++	+	−	Complete
Serum 9	+	+	±	Partial
Serum 10	−	+	+	−
Serum 11	−	+	+	−
Serum 12	+	+	+	−
Serum 13	+	±	±	−
Serum 14	−	±	±	−
Serum 15	−	±	±	−

^*Results are scored for intensity of the 220-kd band: ++ = very strong band; + = clear band; ± = faint band; − = no band. Autoradiograms were exposed for 18 hours. ³⁵S-methionine IPP = radioimmunoprecipitation results with ³⁵S-methionine–labeled extract of HeLa cells (as shown in Figure 2); CTD18 = carboxyl terminal domain 18; MAb = monoclonal antibody.

From a structural perspective, it should be noted that 5 sera that recognized the 145-kd polypeptide of Pol II in immunoblots were unable to immunoprecipitate this enzyme in the presence of CTD 18. Thus, antibodies to the 145-kd polypeptide alone appear to be unable to immunoprecipitate the intact holoenzyme. The simplest explanation is that the relevant epitopes on the 145-kd polypeptide are inaccessible in the assembled complex.

DISCUSSION

Pol II is a complex assembly of at least 10 polypeptides that are held together through noncovalent forces (7,8). Because this complex is stable in IPP buffer, antibodies that bind any of its subunits should theoretically immunoprecipitate the entire structure under the conditions used in the present study. We used immunoblotting and immune inhibition assays to evaluate the Tyr-Ser-Pro-Thr-Ser-Pro-Ser repeat of the Pol II CTD as an autoantigen. This hydrophilic region is thought to extend in a linear manner and to interact with basal transcription factors that assemble around TATA boxes. In one model, phosphorylation of CTD adds a negative charge and reduces its interaction with TFIID, thus permitting forward movement of the polymerase as transcription is initiated (26).

The present results demonstrate that the CTD is a dominant target for many scleroderma sera that contain antibodies to Pol II. Some sera lose their ability to immunoprecipitate the large Pol II subunit when this region is excised (5) or when inhibited with CTD18. In addition, it should be noted that mouse MAb to the large Pol II subunit are specific for the CTD and that the positive control MAb 8wgl6 is indistinguishable from the 5 CTD-specific patient sera in terms of immunoprecipitation, immunoblotting, and inhibition with the CTD18 peptide (15,27). Thus, the CTD appears to provide a dominant immunogenic epitope in both mice and humans.

These results relate directly to recent observations by Satoh et al (6). Those investigators observed antibodies to Pol II in sera of patients with SLE and overlap syndromes as well as those with scleroderma. Scleroderma was characterized by antibodies to all 3 polymerases, as observed in the present study, and they noted that anti–Pol II antibodies in scleroderma sera were mainly specific for the phosphorylated form of this enzyme. The methods used in our study do not permit us to distinguish relative specificities for the 2 forms of the polymerase since there is no way to control for spontaneous phosphorylation or dephosphorylation in the immunoprecipitation assays. As shown in Figure 2, CTD 18 inhibits immunoprecipitation of both the phosphorylated (240-kd) and nonphosphorylated (220-kd) versions of the large Pol II subunit. It is possible that the sera used in our study contained antibodies to both phosphorylated and nonphosphorylated CTD forms and that inhibition is achieved because the synthetic peptide is partially phosphorylated by cell extracts during immunoprecipitation. Alternatively, the immunoprecipitation assay used to detect autoantibodies to Pol II may have preselected for sera directed mainly against the nonphosphorylated form of this enzyme.

Several factors probably contribute to the interactions of the CTD with the immune system. It has a high content of hydrophilic amino acids and a carboxyl terminal location that is likely to be exposed to the solvent at the surface of the enzyme. Parenthetically, it should be recalled that a carboxyl location (28,29) and charged residues (30) are features of a number of autoantigenic epitopes. In addition, its repetitive sequence may also contribute to its ability to activate B lymphocytes (29,31–34). Thus, it would not be surprising if the CTD were uniquely predisposed to function as an autoantigen.

These observations suggest a hierarchy of linked antibodies to the RNA polymerases in scleroderma. Autoantibodies to Pol I and Pol III appear to be the most common, and in most instances are found simultaneously in individual sera. In some patients, this response seems to expand to include Pol II (5). Indeed, this hierarchy may envelop transcription proteins generally. Scleroderma has long been associated with autoantibodies to nucleoli, the nuclear area most active in transcription, and with antibodies to DNA topoisomerase I, a regulator of transcription (35). Recently, we noted the coordinated expression of antibodies to all of the polypeptide subunits of DNA-dependent protein kinase (Ku protein and p350) in patients with scleroderma (36). This enzyme phosphorylates a number of transcription factors, including the CTD of Pol II. Thus, it is seems that much of the humoral autoimmune response associated with scleroderma is directed against nuclear proteins involved with transcription. Further knowledge of how this focal point for autoimmunity comes about should provide important insights into the etiology and pathogenesis of scleroderma.