A Human Autoantibody Specific for a Unique Conserved Region of 28 S Ribosomal RNA Inhibits the Interaction of Elongation Factors la! and 2 with Ribosomes*

An autoantibody reactive with a conserved sequence of 28 S rRNA (anti-28 S) was identified in serum from a patient with systemic lupus erythematosus. Anti-28 S protected a unique 59-nucleotide fragment synthesized in vitro against RNase T1 digestion. RNA se- quence analysis revealed that it corresponded to residues 1944-2002 in human 28 S rRNA and 1767-1825 in mouse 28 S rRNA. These sequences are identical and highly conserved throughout all known eukaryotic 28 S rRNAs. In addition, this fragment is homologous to residues 1052-1 110 of Escherichia coli 23 S rRNA that lies within the GTP hydrolysis center of the 50 S ribosomal subunit. Anti-28 S and its Fab fragments strongly inhibited poly(U)-directed polyphenylalanine synthesis, but had no effect on ribosomal peptidyltransferase activity. This effect resulted from inhibi- tion of the binding of elongation factors EF-la and EF-2 to ribosomes and of the associated GTP hydrolysis. The inhibitory effect was almost completely suppressed by preincubation of anti-28 S with 28 S rRNA or in vitro synthesized RNA fragments containing the immunoreactive region. These results show that the immunoreactive conserved region of 28 S 50 "C). The ribosome-EF-2 complexes were formed in a solution containing 30 pmol of 80 S ribosomes, 100 pmol of [I4C]EF-2 (120 cpm/ pmol), 0.1 mM GuoPP(NH)P, 5 mM MgCl?, 50 mM KCl, 0.2 mM dithiothreitol, and 20 mM Tris-HC1, pH 7.5. Complex formation was analyzed as described for EF-la.

that lies within the GTP hydrolysis center of the 50 S ribosomal subunit. Anti-28 S and its Fab fragments strongly inhibited poly(U)-directed polyphenylalanine synthesis, but had no effect on ribosomal peptidyltransferase activity. This effect resulted from inhibition of the binding of elongation factors EF-la and EF-2 to ribosomes and of the associated GTP hydrolysis. The inhibitory effect was almost completely suppressed by preincubation of anti-28 S with 28 S rRNA or in vitro synthesized RNA fragments containing the immunoreactive region. These results show that the immunoreactive conserved region of 28 S rRNA participates in the interaction of ribosomes with the two elongation factors in protein synthesis.
Interaction of the two elongation factors EF-Tu and EF-G with ribosomes is required for the protein synthesis-elongation cycle in prokaryotes: EF-Tu transports the codon-specific aminoacyl-tRNA to ribosomes, and EF-G promotes translocation of newly synthesized peptidyl-tRNA from the aminoacyl-tRNA binding site (A site) to the peptidyl-tRNA binding site (P site) together with its associated mRNA (Kaziro, 1978;Weissbach, 1980). The action of each elongation factor is accompanied by GTP hydrolysis. Current evidence from genetic and biochemical studies suggests that Escherichia coli * This work was supported by a grants-in-aid for Scientific Research (63770173) from the Ministry, Education and Culture of Japan and a fund from the Japan Society for the Promotion of Science for the Japan-United States Cooperative Science Program. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 4 To whom correspondence and reprint requests should be ad-  (Skold, 1983;Noller, 1984;Hausner et al., 1987;Thompson et al., 1988;Moazed et al., 1988;Dahlberg, 1989). The 1067 region is the binding site of the antibiotic thiostrepton, a powerful inhibitor of ribosome-associated GTP hydrolysis activities (Thompson et al., 1982;Cundliffe, 1986). The 2660 region is the universally conserved loop called the "a-sarcin/ricin domain" which is the site of action of the cytotoxins a-sarcin and ricin; a-sarcin cleaves the G2661-A2662 phosphodiester linkage and ricin removes adenine 2660 in an N-glycosidase reaction (Endo and Wool, 1982;Endo et al., 1987;Hausner et al., 1987).
In the eukaryotic elongation cycle, EF-la and EF-2 serve as the counterparts of E. coli EF-Tu and EF-G, respectively (Weissbach, 1980). The elongation factor-dependent functions of eukaryotic ribosomes are inhibited by a-sarcin and ricin that modify the "a-sarcinlricin domain" of 28 S rRNA (Benson et al., 1975;Montanaro et al., 1975;Fernandez-Puentes and Vazquez, 1977). The results imply that 28 S rRNA plays an important role in the interaction of elongation factors with the eukaryotic ribosome.
Antibodies have been successfully used for determining the functional contributions of specific epitopes within E. coli ribosomal protein L2 (Nag et al., 1986) and eukaryotic ribosomal P proteins (Uchiumi et al., 1990). Although RNA is a poor immunogen for experimental animals, certain patients with autoimmune diseases produce RNA-reactive antibodies that recognize tRNAA'" (Bunn et al., 1986), tRNAMet (Wilusz and Keene, 1986), U1 RNA (Wilusz and Keene, 1986), and rRNA (Lamon and Bennett, 1970;Hardin et al., 1982;Wilusz and Keene, 1986). It has not been clear, however, whether these RNA-reactive antibodies recognize specific sites on the RNA molecules. Recently, Deutscher and Keene (1988) found a unique autoantibody that recognizes a conformational epitope in one of the stem-loop structures of U1 RNA. This finding suggests the possibility that anti-RNA antibodies could be used to specifically map the function of the immunoreactive region of the RNA molecule.
In this paper, we show a new approach for studying the function of eukaryotic rRNA using a site-specific RNA autoantibody. We have detected autoantibodies specific for 28 S rRNA in the sera of patients with systemic lupus erythematosus (SLE),' and have identified a single immunoreactive region within 28 S rRNA using one of these patients' sera. This region is highly conserved in all eukaryotes and homologous to the region around position 1067 in domain I1 of E. coli 23 S rRNA. We have also tested the effect of the antibody on ribosome function and have shown that the antibody inhibits the binding of elongation factors EF-la and EF-2 as well as their ribosome-associated GTP hydrolysis. These results lead to the conclusion that the immunoreactive region of 28 S rRNA plays a fundamental role in the interaction of ribosomes with elongation factors in protein synthesis.

EXPERIMENTAL PROCEDURES
Ribosomes, Ribosomal RNA, and Ribosomal Protein-High saltwashed ribosomes were prepared from HeLa S3 cells (Horak and Schiffmann, 1977), rat liver (Uchiumi et al., 1986), and Artemia salina (Iwasaki and Kaziro, 1979). The 40 S and 60 S ribosomal subunits were isolated from rat ribosomes by sucrose gradient centrifugation, as described previously (Uchiumi et al., 1987). The ribosomal subunits of E. coli ribosomes were prepared by zonal centrifugation, as described previously (Kenny et al., 1979). Ribosomal RNAs were extracted from ribosomes with phenol/chloroform. The 28 S, 18 S, and 5 S RNAs were fractionated by sucrose gradient centrifugation (Uchiumi et al., 1983), and each RNA was isolated. The total 80 S protein was obtained by extraction of ribosomes in 66% acetic acid, 33 mM MgC12, and concentrated by precipitation with 7 volumes of cold acetone. A protein fraction enriched with the three acidic phosphoproteins PO, P1, and P2 (P proteins) was obtained from rat 80 S ribosomal proteins by elution from a CM-cellulose column equilibrated with 6 M urea, 20 mM Tris-HC1, pH 7.8. The eluted proteins were dialyzed against 0.3 M KCl, 5 mM 2-mercaptoethanol, 20 mM Tris-HC1, pH 7.5.
I n Vitro Synthesis of RNA Fragments-The three DNA fragments, SalI/EcoRI, BglIIIEcoRI, and EcoRIISmaI (see Fig. 3A) spanning the 28 S rRNA gene were obtained from a ribosomal DNA clone that had been isolated from mouse genomic DNA (Kominami and Muramatsu, 1987). The DNA fragment for the H1 fragment (BamI/BglII) was isolated from a cosmid clone encompassing a whole repeat unit of human ribosomal RNA gene. The DNA fragments were subcloned into the corresponding restriction sites of an SP-6 vector, pSPT-18 ized by BamHI (Ml) and BglII (M2), EcoRI (M3), SmaI (M4), and (Boehringer Mannheim). These recombinant plasmids were linear-BglII (Hl), respectively (see Fig. 3A) and used in the transcription reaction. The reaction mixture (500 pl) contained 6 mM MgC12, 10 mM NaCl, 2 mM spermidine, 1 mM dithiothreitol, 10 pg of each template DNA, 200 units of SP-6 polymerase (Takara), and 500 p~ concentration each of ATP, GTP, CTP, and UTP supplemented with 25 pCi of [a-32P]UTP at 37 "C for 90 min. The transcripts were purified by gel filtration on a Sephadex G-50 (Pharmacia LKB Biotechnology Inc.) column. For the short fragment H2 corresponding t o residues 1922-2020 of human 28 S rRNA, the corresponding DNA was obtained from human DNA using the polymerase chain reaction (Saiki et al., 1988). The two primers (residues 1922-1941 and 2001-2020) containing the HindIII and XbaI restriction sites at the 5'ends, respectively, were chemically synthesized on an Applied Biosystems synthesizer. The amplified DNA fragments were cut with HindIII and XbaI, and the resultant fragments were cloned into pSPT-18. The plasmid was linearized with XbaI and transcribed as described above. The transcripts were purified with Sephadex G-50.
Autoantibodies-Sera from patients with systemic lupus erythematosus were obtained from The Hospital for Special Surgery, Cor-ne11 University Medical Center, New York and Dept. of Medicine (II), Niigata University School of Medicine, Niigata. The IgG was purified from each serum on a protein A-agarose (Bio-Rad) column. The Fab fragments were prepared by papain digestion (Porter, 1959). The Fc fragments and undigested IgG were removed with protein Aagarose. Homogeneity of IgG and Fah was analyzed by 16% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Laemmli, 1970).
Enzyme-linked Immunosolvent Assay-Enzyme-linked immunosolvent assay was performed as described by Engvall (1980). Briefly, antigenic samples of ribosomal subunits and proteins were adsorbed to the wells of an enzyme immunoassay/radioimmunoassay plate (Costar) by incubation a t 4 "C overnight. After blocking the surface of the wells by incubation with 2% bovine serum albumin in phosphate-buffered saline, serum samples diluted to various extents were added into the well, and the plates were incubated for 2 h a t room temperature. Then, alkaline phosphatase-conjugated goat F(ab'), anti-human IgG (Tago) was added to each well and incubated for 2 h a t room temperature. The chromogenic reaction was performed using the Alkaline Phosphatase Substrate Kit (Bio-Rad).
Labeling of the 60 S Subunits-Rat 60 S ribosomal subunits were labeled at the 3'-ends of RNA constituents using [5'-32P]pCp and T4 RNA ligase (England et al., 1980) and purified on a Sephadex G-150 column.
Immunoprecipitation-Protein A-purified SLE anti-28 S IgG (15 pg/sample) was incubated with each of the RNA samples or 60 S ribosomal subunits in 100 pl of buffer A (100 mM KC1, 5 mM MgC12, and 20 mM Tris-HC1, pH 7.5) at 30 "C for 10 min. Each sample was mixed with another 100 p1 of buffer A containing 1 mg of protein A-Sepharose (Pharmacia LKB Biotechnology Inc.) and incubated with mixing a t 4 "C for 1 h. The antibody-bound beads were washed four times with 500 p1 of buffer A. The bound RNA was recovered by extraction with phenol/chloroform (1:l) and ethanol precipitation.
Ribonuclease Protection Assay-The purified RNA samples (1-2.5 pg) were incubated with 15 pg of the anti-28 S IgG in 100 p1 of buffer A a t 30 "C for 10 min in the presence of 25 pg of E. coli tRNA and then digested with 2 units of RNase T1 (Sankyo) a t 30 "C for 20 min.
The protected RNA fragments were recovered by immunoprecipitation with protein A-Sepharose as described above.
RNA Sequencing-The antibody-protected RNA fragments were labeled a t their 5' ends to high specific activity with [y3'P]ATP (5 mCi/pmol, Amersham) and T4 polynucleotide kinase (Takara) as described by Lillehaug and Kleppe (1975). The labeled fragments were purified by 12% polyacrylamide, 8 M urea gel electrophoresis, and their sequences were determined by partial enzymatic digestion with RNase T1, RNase U2 and Phy M and 12% polyacrylamide, 8 M urea gel electrophoresis (Donis-Keller, 1980).
Elongation Factors-Elongation factors EF-la and EF-2 were purified from pig liver according to Iwasaki and Kaziro (1979).

RESULTS
Characterization of Anti-28 s Serum-We surveyed sera from SLE patients for the presence of anti-ribosomal antibodies by enzyme-linked immunosorbent assay. Several sera showed specificity for ribosomal 60 S subunits. As shown in Fig. 1, a serum was highly reactive both with 80 S ribosomes from HeLa cells and ribosomal 60 S subunits from rat liver, but not with total proteins extracted from HeLa cell ribosomes. The serum showed very low reactivities with rat 40 S subunits, E. coli ribosomal 30 S and 50 S subunits. These results suggested that this serum contained an antibody(ies) reactive with rRNAs in the 60 S subunit. To confirm this, we attempted to inhibit binding of the antibody to the 60 S subunit using each of the isolated rRNA components. As shown in Table I, preincubation of the IgG from this serum with 28 S rRNA prevented the immunoprecipitation of the 60 S subunits, while preincubation with the other RNA molecules had little effect. Since this serum also possessed low titer of activity against the three acidic phosphoproteins PO, P1, and P2 (P proteins) (data not shown), we tested for the effect of preabsorption of the IgG with a mixture of the partially purified three P proteins. No effect was observed in this assay (Table I), suggesting that the serum contained negligible amount of anti-P antibodies compared to the anti-28 S rRNA activity (anti-28 s). These results were consistent with the results by enzyme-linked immunosorbent assay shown in Fig. 1. Fig. 2 shows that the IgG from this serum immunoprecipitated 28 S rRNAs from rat liver and HeLa cells, but did not immunoprecipitate 18 S rRNAs. These results indicate that the anti-ribosome activity of the SLE serum used here is predominantly due to the anti-28 S antibody.
Determination of the Immunoreactive Region within 28 S rRNA-The finding that the anti-28 S antibody reacted to a similar extent with 28 S rRNAs from HeLa cells and rat liver Reactivities of the anti-28 S serum with ribosomal subunits. HeLa cell 80 S ribosomes and the small and large subunits obtained from rat liver and E. coli (0.1 A260 unit/well) and total 80 S protein from HeLa cell ribosomes (10 pg/well) were adsorbed to the wells of a microtiter plate. The serum was diluted with 1% bovine serum albumin in phosphate-buffered saline, and IgG binding to each test antigen was determined by enzyme-linked immunosorbent assay. IgG was purified with protein A from a SLE serum that specifically reacts with the 60 S subunits (Fig. 1). The IgG (15 pg) was incubated with various ribosomal components at 30 "C for 10 min prior to mixing with rat 60s ribosomal subunits (0.1 A260 unit: 9400 cpm) that had been labeled at the 3' ends of RNA using [5'-"P]pCp and T4 RNA ligase (England et al., 1980). Immunoprecipitation was performed with protein A-Sepharose beads as described under "Experimental Procedures."
suggested that it might recognize a conserved region of 28 S rRNA, as has been shown for anti-P autoantibody reactive with ribosomal P proteins (Elkon et al., 1986). In order to determine whether the anti-28 S recognizes a single site or multiple sites of protein-free 28 S rRNA, DNA fragments encoding various regions of mouse 28 S rRNA were cloned into the SP-6 vector (Fig. 3A), and the anti-28 S antibody was tested for binding to their RNA transcripts by immunoprecipitation. Of the four different transcripts M1 (SalI/ BamHI), M2 (SalI/BglII), M3 (BglII/EcoRI), and M4 (EcoRI/SmaI), only the M2 transcript was precipitated by the anti-28 S antibody (Fig. 3B). This result shows that the antibody binding site lies only within the region between the BarnHI and BglII sites (residues 1242-2105). The antibody was also reactive with the transcript H1 from the BamHI/ BglII region (residues 1406-2328) of human 28 S rRNA (Fig.  3B).
To further define the immunoreactive region of RNA, we performed an RNase protection assay using the anti-28 S antibody. Each transcript was incubated with the antibody A Site-specific Autoantibody to 28 S rRNA and then digested with RNase T1. The resultant RNA fragments were immunoprecipitated and subjected to polyacrylamide gel electrophoresis (Fig. 3C). A single product of about 60 nucleotides long was precipitated from either transcript M2 or H1. No fragment was precipitated when IgG from normal serum was tested for this protection assay nor when the anti-28 S antibody was added after the digestion (data not shown). The antibody-protected RNA fragments were subjected to RNA sequence analysis. The sequence of the protected fragment from the H1 transcript corresponded to residues 1944-2002 of human 28 S rRNA (Fig. 4A) and that from the M2 mouse transcript corresponded to residues 1767-1825 (data not shown). These sequences are identical with each other and have 58% homology with the region between residues 1052 and 1110 in domain I1 of E. coli 23 S rRNA, which is known to lie within the "GTP hydrolysis domain" of the 50 S ribosomal subunit. A secondary structure of the immunoreactive region is shown in Fig. 4B according to the model of Gorski et al. (1987). The purified antibody-protected RNA fragment could rebind to the antibody in the presence of 5 mM M&12 as the M2 and H1 transcripts bound, but not when Mg2' was removed with EDTA (data not shown). These results indicate that the anti-28 S is reactive with a single region of 28 S rRNA and that this antibody recognizes a Me-induced RNA conformation.
Effect of the Anti-28 S Antibody on Protein Synthesis-Elongation Cycle-To investigate the functional effect of the anti-28 S antibody, poly(U)-directed polyphenylalanine synthesis was attempted in the presence of the antibody. As shown in Fig. 5, A and B, both of the anti-28 S IgG and its Fab fragments completely inhibited the polyphenylalanine synthesis activity of HeLa cell ribosomes. Antibodies from six SLE sera without the anti-28 S activity as well as IgG obtained from normal serum had no effect on polyphenylalanine synthesis. The anti-28 S antibody had similar inhibitory effects on rat liver and A. salina ribosomes (data not shown).
We examined the effects of preabsorption of the antibody with various components of ribosomes or RNA transcripts (Fig. 5C). The inhibitory effect on polyphenylalanine synthesis was fully suppressed by preincubation of the antibody with 28 S rRNA. No relief of the inhibition was given by the other RNAs nor by the P proteins. Among the subfragments of 28 S rRNA, only the M2 and H1 transcripts gave suppression of the inhibition. These results are consistent with those in the immunoprecipitation assay (Fig. 3). To confirm the results of the RNase protection study, we synthesized the shorter transcript H2 which corresponds to residues 1922-2020 of human 28 S rRNA. This H2 transcript was immunoprecipitated by the anti-28 S antibody (data not shown) and also reversed the inhibition of polyphenylalanine synthesis (Fig. 5C). These results indicate that the inhibition of polyphenylalanine synthesis results from antibody binding to a 59-nucleotide-long immunoreactive region of 28 S rRNA (see Fig. 4B). Which Step of the Elongation Cycle Is Inhibited by the Antibody?-The anti-28 S antibody was tested for its effect on individual steps of the protein synthesis-elongation cycle. Fig. 6A shows the effect on peptidyltransferase activity, in which the formation of N-acetlyphenylalanyl-puromycin was assayed using isolated 60 S ribosomal subunits. The antibody had no significant effect on this activity. Fig. 6B shows the effect on ribosome-associated GTP hydrolysis activities dependent on the two elongation factors EF-lo and EF-2. The antibody inhibited both of the factor-dependent GTP hydrolysis activities in a similar manner. The EF-lo-dependent ['4C]phenylalanyl-tRNA binding to 80 S ribosome-poly(U) complexes was assayed by membrane filtration. The anti-28 Fab fragments ( B ) at 37 "C for 10 min and were then tested for polyphenylalanine synthesis, as described under "Experimental Procedures." 0, anti-28 S IgG; A, IgG from normal serum. The 100% activity corresponded to 47 pmol of ["Clphenylalanine polymerized. C, anti-28 S IgG (30 pg) was preincubated with various ribosomal components or RNA fragments (0.7 AZG0 unit of 28 S rRNA, 18 S rRNA, Ml-M4, H1, or yeast tRNA; 0.1 A2w unit of 5 S rRNA or H2; 10 pg of P protein fraction) at 37 "C for 10 min prior to mixing with the 80 S ribosomes. After further incubation at 37 "C for 10 min, polyphenylalanine synthesis was assayed. Each of the RNA fragments used was transcribed as shown in Fig. 3A. The H2 fragment (residues 1922-2020) spanning the immu- S antibody completely blocked its binding (Fig. 6C). The extent of inhibition in the elongation factor-dependent activities is comparable to that observed for polyphenylalanine synthesis, suggesting that the inhibition of the polymerization was due to the impairment of these elongation factor-dependent events. The effect of the antibody on binding of the two elongation factors to ribosomes was analyzed directly by sucrose gradient centrifugation. As shown in Fig. 7, both of the complexes [3H]EF-la. Phe-tRNA. GuoPP(NH)P and ["C] EF-2 GuoPP(NH)P failed to bind to the ribosomes that had been pretreated with the anti-28 S antibody. The sedimentation pattern of the 80 S ribosomes was little affected by the addition of the antibody, suggesting no formation of ribosome dimers or aggregates with the antibody. These results demonstrate that the anti-28 S antibody completely blocks the binding of elongation factors to 80 S ribosomes, and, as a consequence, their function in elongation. The results imply the participation of the immunoreactive region of eukaryotic n"" n , g g g a L 5 1 I I X 2 % r 1 r e 4 c l u r e l n 28 S rRNA in the interactions of ribosomes with both elongation factors EF-la and EF-2.

DISCUSSION
Anti-rRNA antibodies have previously been identified in the sera of patients with autoimmune diseases (Lamon and Bennett, 1970;Hardin et al., 1982;Wilusz and Keene, 1986), but the fine specificity of these antibodies has not been characterized further. The present study demonstrates that certain SLE patients produce autoantibodies reactive with a specific site of 28 S rRNA. One of these sera has been extensively characterized. The immunoreactive RNA region recognized by the anti-28 S antibody was isolated by RNase digestion and immunoprecipitation, and its sequence was determined. The nucleotide sequence is phylogenetically highly conserved in eukaryotes from human to yeast. Furthermore, this segment is 58% homologous to residues 1052-1110 of E. coli 23 S rRNA that is within the GTP hydrolysis center of the 50 S subunit.  6. Effects of the anti-28 S antibody on peptidyltransferase activity (A), on elongation factor-dependent GTPase activities ( B ) , and on EF-la-dependent aminoacyl-tRNA binding (C). A , rat liver 60 S ribosomal subunits (15 pmol) were preincubated with increasing amounts of the anti-28 S IgG at 37 "C Btibody to 28 S rRNA The antibody was employed as a probe for studying the function of this region in 28 S rRNA. The inhibitory effect of the antibody on elongation factor-dependent ribosome functions suggests that the immunoreactive RNA region is involved in the interactions of ribosomes with both EF-la and EF-2. It is unlikely that the inhibition of ribosome function is due to antibody activities other than the anti-28 S. The inhibitory effect of the antibody was completely blocked by preincubation with 28 S rRNA or RNA fragments containing the immunoreactive region, but not at all with the other RNA components. Although the anti-28 S serum used contained low titer of anti-P antibodies as determined by immunoblotting and enzyme-linked immunosorbent assay, anti-P levels were approximately 1/100 lower than anti-60 S subunit levels and were 1/1000 lower than the monoclonal anti-P antibody used previously for inhibition tests (Uchiumi et al., 1990) (data not shown). It is also unlikely that the inhibition by the anti-28 S antibody resulted from a nonspecific steric effect due to formation of antibody-ribosome complexes. First, Fab fragments from the anti-28 S IgG also showed complete inhibition of polyphenylalanine synthesis (Fig. 5B). Second, the antibodies had no effect on peptidyltransferase activity (Fig. 6 A ) . Third, the sedimentation profile of 80 S ribosomes was not altered by the antibody binding (Fig. 7).

The GTP Hydrolysis Domain in Eukaryotic Ribosomes-
The GTP hydrolysis domain, one of the active sites of the ribosome, has been extensively and best characterized in E. coli ribosomes. Protein L7/L12 is a component of this domain and participates in interactions with elongation factors Tu and G and initiation factor 2 (Hamel et al., 1972;Moller, 1974;Weissbach, 1980). The action of each factor is accompanied by GTP hydrolysis, which occurs to only a minimal extent on the L7/L12-deficient ribosomal core particles. Two homodimers of L7/L12 together with protein L10 form a pentameric complex, named the L8 complex (Petterson et al., 1976;Petterson and Liljas, 1979), and are stably integrated in the large ribosomal subunit through L10. The impairment of EF-G-dependent GTP hydrolysis is also observed with ribosomes lacking protein L11 or its homologue in Bacillus megaterium (Stark and Cundliffe, 1979;Stark et al., 1980). Protein L11 was affinity-labeled by a photoactivated GTP analogue (Maassen and Moller, 1978). Both protein L11 and the L8 complex bind to overlapping sequences within residues 1028-1124 of 23 S rRNA (Schmidt et al., 1981;Beauclerk et al., 1984). The function of this RNA portion has been studied with an antibiotic, thiostrepton, a powerful inhibitor of ribosome-associated GTPase activities dependent on EF-G, EF-Tu, and IF-2 . Thiostrepton binds directly to a specific site within residues 1052 and 1110 of 23 S rRNA. Methylation of A1067 or mutation of this base prevents binding of thiostrepton and renders the ribosome insensitive to the drug (Thompson et al., 1982(Thompson et al., , 1988  show that EF-G protects A1067 and A1069 (Moazed et al., 1988). Furthermore, EF-G has been cross-linked to 23 S rRNA within residues 1055-1081 (Skold, 1983). These lines of evidence all indicate that the residues 1052-1110 of 23 S rRNA constitute a functional domain involved in the binding of elongation factors and GTP hydrolysis. The present data provide the first evidence for the presence in eukaryotic 28 S rRNA of a region functionally homologous to the 1052-1110 region of E. coli 23 S rRNA. The results with the E. coli system indicate the association of both RNA and protein elements in the GTP hydrolysis domain. In a previous paper, we showed that the homologous carboxyl-terminal regions of three eukaryotic P proteins PO, P1, and P2 also participate in the binding of EF-la and EF-2 to ribosomes by using monoclonal anti-P antibodies (Uchiumi et al., 1990). Proteins P1/P2 and PO are presumed to be eukaryotic counterparts of E. coli L7/L12 and L10, respectively (Uchiumi et al., 1987). Recent experiments have shown that the P proteins bind to the RNA transcripts H1, H2, and M2 containing the immunoreactive region of 28 S rRNA, but not to any other RNAs lacking this region, as assayed by immunoprecipitation with anti-P monoclonal antibodies.' These findings indicate that the P proteins, probably as a complex, bind to 28 S rRNA at or near the immunoreactive region for anti-28 S . It seems likely, therefore, that this RNA-' T. Uchiumi and R. Kominami, unpublished results.
protein complex forms the eukaryotic counterpart of the GTP hydrolysis domain of E. coli ribosomes. Both anti-28 S (this study) and anti-P (Uchiumi et ai., 1990) antibodies inhibit polyphenylalanine synthesis, elongation factor binding, and factor-dependent GTPase activity, suggesting that both antibodies inhibit the same step of protein synthesis. However, the anti-28 S antibody inhibited all of these activities more completely than a monoclonal anti-P antibody; the anti-P antibody never inhibits more than 80% regardless of antibody excess (see Uchiumi et al., 1990, Fig. 4), whereas the anti-28 S antibody inhibits the factor binding nearly 100%. This may reflect the different functional contribution of the RNA portion and the P proteins in the domain and supports the notion that rRNA plays a primary role in the mechanism of translation whereas ribosomal proteins play an auxiliary role facilitating RNA function by inducing unique RNA conformation. Autoantibodies Specific to RNA Conformation-We have obtained 11 anti-28 S sera from 80 SLE patients that appear to recognize the same site in 28 S rRNA as that described here. Other RNA sites have not yet been identified.3 The reason why this conserved region of 28 S rRNA is predominant for the immune response is not clear. The M e dependence of the antibody-RNA interaction may indicate the role of secondary and tertiary RNA structure, but does not rule T. Uchiumi, T. Sato, and R. Kominami, unpublished results. out the contribution of primary structure. Site-specific antibodies that recognize a conformation within U1 RNA and cruciform DNA structures have been found by other investigators (Frappier et ai, 1987;Deutscher and Keene, 1988).
The etiology of the production of these conformation-specific antibodies remains obscure. Deutscher and Keene (1988) suggested that the anti-U1 RNA antibody arose as an antiidiotype against an autoantibody that recognized the RNA binding domain of a U1 RNA-associated protein. A similar idiotype-anti-idiotype relationship was suggested as an explanation for the coexistence of antibodies to alanyl-tRNA synthetase and tRNAA'" in a myositis patient's serum (Bunn et al., 1986). It is of interest that all of the anti-28 S sera identified by us also contained anti-P activities. However, it seems unlikely that anti-28 S originates as an anti-idiotype against anti-P antibody, since anti-28 S binds to intact ribosomes that have bound P protein. The antibody binding site appears therefore to be close to, but distinct from, that for P proteins. Taking into account our recent observation that the P proteins bind to the immunoreactive RNA fragment described above, it seems more likely that the co-production of anti-28 S and anti-P antibodies is due to the close proximity of these two antigens in the 60 S subunit. Further detailed knowledge about the tertiary and quaternary structure of this antigenic domain of ribosomes could provide insights into not only the selective production of autoantibodies, but also the molecular mechanism of action of the two elongation factors in protein synthesis.