Proc. Natl. Acad. Sci. USA Vol. 92, pp. 12309-12313, December 1995 Biochemistry

UGA suppression by a mutant RNA of the large ribosomal subunit (nonsense suppression/termination defect/23S rRNA, domain II/conserved nucleotide) DAVID K. JEMIOLO*, FRANCES T. PAGEL, AND EMANUEL J. MURGOLAt Department of Molecular Genetics. Box 11, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030 Communicated by John Roth, University of Utah, Salt Lake City, UT, August 16, 1995 (received for review January 25, 1995)

ABSTRACT A role for rRNA in peptide chain termination was indicated several years ago by isolation of a 16S rRNA (small subunit) mutant of Escherichia coli that suppressed UGA mutations. In this paper, we describe another interesting rRNA mutant, selected as a translational suppressor of the chain-terminating mutant trpA(UGA211) of E. coli. The find- ing that it suppresses UGA at two positions in trpA and does not suppress the other two termination codons, UAA and UAG, at the same codon positions (or several missense mutations, including UGG, available at one of the two posi- tions) suggests a defect in UGA-specific termination. The H indI' suppressor mutation was mapped by plasmid fragment ex- changes and in vivo suppression to domain II of the 23S rRNA SaLI gene of the rrnB operon. Sequence analysis revealed a single SphI base change of G to A at residue 1093, an almost universally conserved base in a highly conserved region known to have ;l {\E EcoRI specific interactions with ribosomal proteins, elongation fac- HpaI tor G, tRNA in the A-site, and the peptidyltransferase region SacI of 23S rRNA. Several avenues of action of the suppressor Bgl I I rsp45 I mutation are suggested, including altered interactions with FIG. 1. Map of plasmid pDJ100. Open bars (from HindIlI clock- release factors, ribosomal protein L1I, or 16S rRNA. Regard- wise to BamHI), pBR322; stippled bars, phage A DNA; hatched bar, less of the mechanism, the results indicate that a particular unidentified 300-bp insertion; solid and shaded bars, E. coli chromo- residue in 23S rRNA affects peptide chain termination, spe- somal DNA containing the rrnB operon (solid bars represent coding cifically in decoding of the UGA termination codon. regions for the three rRNAs and tRNAGIu); PL, promoter-operator region from phage A; bent arrow, start site and direction of transcrip- A role for rRNA in peptide chain termination was indicated tion of rRNA genes. Unique restriction sites: Kpn I, Csp45I, Sac I, Hpa I, BamHI; double sites: Sph I, Bgl II. HindlIl is shown only for several years ago by isolation of a 16S rRNA (small subunit) reference; there are two other Hindlll sites in pDJ100, both in the 16S mutant ofEscherichia coli that suppressed UGA mutations and rRNA gene. implicated in particular the universally conserved nucleotide C1054, located in the conserved helix 34 (1). A key role for 23S was moved into our strains by way of the closely linked rRNA (large ribosomal subunit) in peptide bond formation, transposon zad-981::mini-kan. We have previously described translocation, and translational accuracy has been clearly the methods for screening prototrophic derivatives of a Trp- established (refs. 2-5 and references therein). The small auxotroph to distinguish suppressed mutants from revertants ribosomal subunit, on the other hand, has generally been (10) and for comparing growth of suppressor-containing thought to be the subunit involved in codon-specific transla- strains on plates (11). The main plasmids used in this study tional events in both elongation and termination. Indeed, until were as follows. pc1857 was derived from pACYC177 and recently, the majority of rRNA suppressors of nonsense (ter- contains a temperature-sensitive bacteriophage A repressor mination codon) mutations obtained were associated with 16S gene (12). pDJ100 was derived from pBR322 and contains the rRNA (6). In this paper, however, we describe a 23S rRNA wild-type rRNA operon, rrnB, under the control of the A PL nonsense suppressor that works at UGA mutant codons in trpA promoter. It is a spontaneous derivative of pNO2680 (13) and but not at UAA or UAG mutations at the same codon contains an '300-bp insertion in the PL region between the positions. A preliminary report of these results was presented two Bgl II sites of pNO2680 (unpublished observations; Fig. 1). at the Cold Spring Harbor Laboratory meeting on the Mo- pDJ101 is the original rrlB(SuUGA)-containing plasmid re- lecular Genetics of Bacteria and Phages, August 18-23, 1992. ported in this study (that is, pDJ100 with the new suppressor mutation). pDJ101[4a] is pDJ101 with an uncharacterized MATERIALS AND METHODS secondary mutation, outside of rrnB, that was found in this The bacterial strains were derived from E. coli K-12. The study to affect plasmid copy number or level of transcription nomenclature for trpA mutations and suppressor genes has from PL. Plasmid DNA preparation and transformations were been described (7). Specific strains are described at first carried out essentially as described by Sambrook et al. (14). mention in Results. Some strains contained a pcnB mutation, Plasmid mutagenesis was performed in vivo. The plasmid was which reduces colEl-type plasmids to a few copies per cell (8, 9). pcnBl, obtained from J. S. Parkinson in strain RP7947 (9), Abbreviations: AmpR, ampicillin resistance (resistant); Ind, indole; 5MT, 5-methyl-DL-tryptophan; EF, elongation factor; RF, release factor. The publication costs of this article were defrayed in part by page charge *Permanent address: Department of Biology, Vassar College, Pough- payment. This article must therefore be hereby marked "advertisement" in keepsie, NY 12601. accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed.

Downloaded by guest on September 29, 2021 12309 12310 Biochemistry: Jemiolo et al. Proc. Natl. Acad. Sci. USA 92 (1995)

introduced into a strain containing the mutator mutDS (15) on functional and transcription of the mutant rrnB increases. In supplemented minimal medium; the strain was then grown in this way, we demonstrated that at 42°C, the suppressor muta- L broth plus thymidine to allow the mutator to work. The tion, present in pDJ101, is lethal to the cell. In the second mutD5-exposed plasmid was harvested and used to transform instance, we used a plasmid preparation of the suppressor various trpA mutant strains to obtain plasmid-associated sup- mutant to transform pairs of isogenic strains containing each pressors. Media and other genetic procedures have been offour trpA UGA mutations. Each pair contained a PcnB+ and described elsewhere (7, 16) or are presented in Results. Most a PcnB- member. In both members of each pair, the rrnB biochemical procedures were performed essentially according operon was maximally transcribed. They differed, however, in to ref. 14. For DNA sequence analyses, small fragments from their plasmid copy numbers by virtue of the status of thepcnB the mutant rrnB operon were separated by agarose gel elec- allele. Table 1 highlights the UGA suppression ability of the trophoresis, purified using the Prep-A-Gene DNA purification suppressor (in pDJ101) and indicates the suppressor's lethality kit (Bio-Rad), and recloned into a vector containing multiple at higher copy number. cloning sites bracketed by T3 and T7 promoters. Plasmid Transformation of PcnB+ strains with pDJ101 yielded ex- preparations were analyzed by automated sequence analysis tremely sick looking, nonviable "colonies." Among them, using T3 and T7 complementary primers, Applied Biosystems however, were a few fast-growing, healthy colonies, so several reaction protocols, and a model 373A DNA sequencer. were purified and examined for a secondary mutation in the plasmid. In two cases, back transformation demonstrated that RESULTS a secondary mutation accounting for healthy growth was plasmid associated. Fragment exchanges between one of the Suppressor Selection. mutD5-treated pDJ100 (see Fig. 1) nondetrimental plasmids, pDJ101[4a], and wild-type pDJ100, was used to transform strain MDA6655 to ampicillin resistance followed by transformations of PcnB+ and PcnB- strains, on minimal containing 50 (AmpR) and Trp+ glucose medium indicated that the secondary, remedial mutation was not in the of per ml Amp-50). MDA6655 is Trp- due to Ag Amp (Min rRNA operon (between Kpn I and, counterclockwise, a UGA nonsense mutation at codon 211 of the trpA gene. It BamHI). Presumably, it lowers the effectiveness of the sup- also contains apcnB mutation, which lowers the copy number pressor (Table 1), and hence its detrimentality, either by of colEl-type plasmids (8, 9), to allow us to obtain rRNA copy number or by lowering transcription mutations that, when highly expressed, might be detrimental to decreasing plasmid from the PL promoter. gene was found the cell. After single-colony isolations, the trpA and Nucleotide Analysis of to still contain the original UGA mutation, indicating that the Molecular Mapping Sequence the Suppressor Mutation. To localize the suppressor mutation Trp+ phenotype was due to suppression. Plasmid extracted to a convenient for DNA sequence analysis and to from the suppressor strain was used to retransform the original fragment recipient strain (MDA6655) to AmpR. All the AmpR trans- exclude the possibility that secondary mutations outside the were in we cut both the wild-type formants were Trp+ (i.e., suppressor positive), indicating that region involved suppression, with appro- the new suppressor was plasmid associated (and presumably in plasmid and the suppressor-containing plasmid the rrnB operon). priate restriction enzymes (see Fig. 1), exchanged the frag- High Expression Detrimentality and Its Reversal by a ments, religated them, and transformed appropriate cells for Vector Mutation. Overproduction of a mutant rRNA encoded the functional test (UGA suppression). The results of several by a cloned rrn operon may be detrimental to cell growth or pairwise cuts established that the suppressor was in the gene even result in lethality. Since the new suppressor rRNA for 23S rRNA, within a 623-bp segment from the Hpa I site to mutation was obtained in a PcnB- strain and in a cloned rrnB the nearer Sph I site. That region comprises essentially all of under the control of the A PL promoter, we were able to test "domain II" (of 6 domains) of 23S rRNA. For DNA sequence the possible detrimentality of the mutant in two ways. One was analysis, we prepared two fragments that constitute the Hpa to conditionally increase the transcription of rrnB in a PcnB+ I/Sal I (counterclockwise) segment of rrnB: the first fragment strain; the other was to test the mutant rrnB plasmid in PcnB+ (238 bp) was from Hpa I/EcoRI and the second (500 bp) was vs. PcnB- strains. In the first instance, we introduced the from EcoRI/Sal I (Fig. 1). Each fragment was cloned into the mutant pDJ101 into a strain containing pcI857 at 30°C. The A multicloning site of a sequencing vector and sequenced in both repressor gene (cI857) encodes a temperature-sensitive repres- directions by an automated DNA sequencer. The only se- sor that is functional at 30°C and so allows little if any quence alteration detectable was in the 500-bp fragment, transcription of the mutant rrnB from the PL promoter. As the between EcoRI and Sal I, a G to A change at residue 1093 (see temperature is raised, however, the repressor becomes less Fig. 2). Table 1. Suppression of trpA UGA mutations at four codon positions UGA codon position 15 211 234 243 Plasmid PcnB- PcnB+ PcnB- PcnB+ PcnB- PcnB+ PcnB- PcnB+ pDJ100 ------pDJ101 - ND + ND ND + + ND pDJ101[4a] - - +- ++ - - + +++ pFP100 ------Patch growth on agar plates was used to compare the growth of strains harboring suppressor-containing plasmids with strains harboring suppressor-negative plasmids (modified from ref. 11). The plates contained glucose minimal medium with Amp at 50 ,g/ml (Min Amp-50) for PcnB- strains or at 100 jig/ml (Min Amp-100) for PcnB+ strains. -, No growth over suppressor-negative strain, up to 6 days; +, + +, and + + +, increasing amounts ofgrowth over control plasmid (= suppression), normalized to growth of each strain on Min Amp with Ind (suppression of the trpA mutation not needed for growth). PcnB-, strain contains apcnB mutation, which reduces ColEl-type plasmids to a few copies per cell (8,9); PcnB+, strain contains the wild-type pcnB gene. ND, not determinable because of lethality of the original suppressor-containing plasmid in pcnB+ strains. Downloaded by guest on September 29, 2021 Biochemistry: Jemiolo et al. Proc. Natl. Acad. Sci. USA 92 (1995) 12311

- - - readthrough by a suppressor is manifested as a zone smaller A1093 than the control strain (suppressor free), whereas a zone of the A 1080 UAA + same size indicates no suppressor-induced readthrough. In this G C A U A AA -A G C C A C A U GA G C A A way, we observed readthrough by A1093 of UGA but not of *AII II I I I I UAA or UAG, while observing readthrough by C517 at all C U C G IA* A U G U G A UA three termination codons. This result parallels the growth G-C 1100 observations on glucose minimal medium, where growth re- A-U quires production not only of full-length trpA-encoded protein C-G but also of catalytically active protein. 105 C-G DISCUSSION IG G- IA A Translational suppression is a most effective way to examine C C C-G A the structure, function, and interactions of any translational G.U macromolecule, as long as and to the extent that that molecule L - is involved in the specificity or accuracy of translation (see ref. A G 23 for general review). Most of the suppressors studied over A L G__ I G-C the past 30 years or so have been tRNAs. However, nonsense suppressor mutations in particular have been found in the G-C genes for release factors (RFs) RF1 and RF2 (24-26), elon- G *U gation factor (EF) EF-Tu (27), and ribosomal proteins (28,29). U OG- 1120 Translational suppressors that mapped to or near rRNA G-C genes in yeast mitochondria were known since the early 1980s (30-32), but it was not until 1989 that the DNA sequence of FIG. 2. A secondary structure representation of a portion (nt one of them was reported (33). The mutant site, in the small 1034-1121) of domain II of E. coli 23S rRNA (adapted from ref. 17). ribosomal subunit RNA, was analogous to residue 517 inE. coli Tick marks are located every 10 nt; numbering is from the 5' terminus of the mature 23S rRNA. The A for G1093 base substitution is the 16S rRNA, mutation of which was later shown to affect mutational change that was found in the suppressor of trpA(UGA2i1) translational accuracy (22). A role for 16S rRNA in peptide described in this report. A (dashed enclosure), the apparent Lii chain termination had been indicated by the 1988 report of an binding region (adapted from refs. 18 and 19); E (dashed enclosure), rrsB mutant that suppresses UGA mutations in trpA, but not the apparent L10.(Ll2)4 binding region (adapted from ref. 18); 0, UAA, UAG, or missense mutations (1). That finding and nucleotides protected by EF-G (20); *, nucleotides protected by subsequent work (6, 34-36), implicated in peptide chain A-site-bound tRNA (21). termination the universally conserved residue C1054 and the secondary structure in which it is located, helix 34. In this Suppressor Specificity. To determine how specific the new paper, we describe a mutation in a cloned E. coli gene for 23S suppressor was, we introduced plasmid pDJ101[4a] into iso- rRNA (the large RNA of the large ribosomal subunit), a genic PcnB+ strains differing only in the nature of the trpA mutation that implicates the large subunit and 23S rRNA in mutation. After purification of the AmpR transformants, sev- particular in decoding-that is, in the codon recognition step eral from each were tested for Trp+. The results were that the of termination of polypeptide synthesis at UGA codons. G1093 to A change in 23S rRNA suppressed UGA mutations In general, an rRNA mutation could lead to suppression of at two trpA positions but not UAA or UAG at the same codon a chain-terminating mutation (readthrough) by either increas- positions (or several missense mutations, including UGG, ing termination codon mistranslation or decreasing the effk - available at one of the two positions). (The -1 frameshift tiveness of the termination mechanism. In the first case a mutation trpE91 was also tested and found not to be sup- decrease in translational accuracy should be relatively i- pressed.) To verify that the trpA nonsense mutants that we specific and may be expected to exhibit suppression of all ti used can indeed reveal nonspecific nonsense suppressors, we types of nonsense (termination) mutation and some missv examined them in the presence of the 16S rRNA mutant C517, mutations. The second should be relatively specific, eith( which suppresses all three termination codons in lacI-Z fu- nonsense mutations (all three termination codons) vs. sions (22). With CS17, we observed suppression, on glucose sense mutations or for one of the three termination codc . minimal medium, of UGA at three trpA positions and of UAA missense mutations and the other termination codons. RL '11s and UAG at two of the three (K. A. Hijazi, F.T.P., and E.J.M., of suppressor-specificity tests suggest that the A1093 suppres- unpublished results). sor is specific for UGA mutations. The trpA UAA and UAG In another approach to the codon specificity of A1093, we mutants used are capable of exhibiting readthrough, on glu- examined the growth of trpA nonsense mutant strains con- cose minimal medium, in the presence of an rRNA mutant that taining A1093 on a plasmid by a modification of the zone of is not codon specific (namely, the 16S mutant designated C517; inhibition procedure (16). In this case, the agar medium ref. 22). Furthermore, in a readthrough test that does not contained a low concentration (1.5 gg/ml) of indole (Ind). On depend on catalytic activity of the trpA-encoded protein, the disk was deposited 1 gg of the trp corepressor 5-methyl- A1093 exhibited readthrough at UGA but not UAA and UAG, DL-tryptophan (5MT). The amount of 5MT to be placed on the while C517 exhibited readthrough at all three codons. Finally, disk was chosen, from several tested, as an amount that gave using trpA UAG mutants, we recently isolated 16S rRNA a small but readable zone with a positive control strain. Each mutants that suppress UAG codons as well as UGA, and one suppressor-free strain contained a particular trpA nonsense that suppresses all three codons (36). The results suggest mutation and the plasmid carrying the wild-type rrnB operon. therefore that A1093 is not an altered accuracy mutant but Media with low Ind and high 5MT (Ind-5MT) can distinguish rather that it causes a defect in UGA-specific termination. premature termination events at nonsense codons in trpA from The A1093 mutation is in the 5' half of 23S rRNA in domain readthrough events that yield a full protein, regardless of the II of six domains (Fig. 2). The immediate neighborhood is a amino acid inserted at the nonsense codon (see ref. 23 for highly conserved region and G1093 itself is extremely well review and original references). Hence growth on Ind-5MT conserved. The finding of this mutation in a selection for indicates readthrough, while failure to grow indicates termi- altered translation further supports the view that phylogeneti- nation. In the zone of inhibition version of the test, cally conserved regions of rRNA play indispensable functional Downloaded by guest on September 29, 2021 12312 Biochemistry: Jemiolo et al. Proc. Natl. Acad. Sci. USA 92 (1995) roles (37). The 1093-containing region, from nt 1038 to nt alteration of such an intersubunit RNA-RNA interaction 1117, contains interaction sites for several proteins: EF-G (20), could produce the suppressor phenotype directly or in con- which has been cross-linked to this region (38); ribosomal junction with RF2 or LI 1. It is interesting that ribosome-bound protein Lii (39), which has been implicated in a number of mRNA has been cross-linked not only to RF2 (50) but also to functions associated with protein synthesis, including termi- both the Lu1-binding region of 23S rRNA (51) and helix 34 of nation; and a pentameric ribosomal protein complex com- 16S rRNA (52). The A1093 mutation could also influence the posed of one LIO and four L12 molecules (40). This region is activity of 16S rRNA indirectly by causing improper subunit involved, along with domain VI, in EF-G-dependent GTPase association. activity and binds the antibiotic thiostrepton, which inhibits Two recent sets of findings from our laboratory should help EF-G-associated GTPase activity (41) and has been shown to elucidate the mechanism of action of A1093. First, after inhibit RF binding to the ribosome in the presence of stop segment-directed PCR random mutagenesis and subcloning of codons (42). tRNA bound to the A-site of the large ribosomal the mutagenized fragment, AmpR transformants of trpA UGA subunit protects residues in this region (nt 1068 and 1071) from mutant strains were screened for plasmid- and fragment- chemical attack (21), and evidence has been presented for associated UGA-specific suppressors. In addition to II inde- interaction of domain II with domain V, which contains the pendent isolates of A1093, suppressors were also found with peptidyltransferase center (43, 44). deletions of G1093, A1095, or U1097, or a base substitution at Several avenues of action can be suggested for the A1093 position 1094, all in the highly conserved hexanucleotide loop mutation-that is, ways in which it might cause a defect from nt 1093 to nt 1098 (ref. 36; W. Xu and E.J.M., unpub- specifically in UGA-directed peptide chain termination. First, lished results). Second, the apparent high-expression lethality it may directly affect RF2, the release factor that works caused by A1093 when cloned into the original pN02680 (13) specifically at UGA termination codons. This RF has several has been found to be high-temperature conditional (36). That points of close contact with both subunits of the ribosome is, when fully expressed from plasmid pN02680 but tested at when it binds in response to a termination codon in the temperatures below 36°C, A1093 is not lethal. Selections for decoding site (45). Furthermore, since it seems that RF3 works 23S rRNA segment-directed PCR mutations that reverse the predominantly with RF2 at UGA-containing stop signals (46), lethality of A1093 at 37°C or 41°C have revealed mutations in the A1093 phenotype may be achieved primarily by altered domain V or domain VI (ref. 36; A. L. Arkov and E.J.M., interaction with RF3. Another possibility rests on the sug- unpublished results). gested role of peptidyltransferase not only in peptide bond Regardless of the mechanism of action of A1093, our results formation but also in hydrolysis of ribosome-bound peptidyl- demonstrate that 23S rRNA, and specifically a particular tRNA. Domain II may interact with domain V, which contains conserved nucleotide, plays a role in decoding the UGA the peptidyltransferase center, leaving open the possibility that termination codon. A1093 decreases the efficiency of the proposed hydrolytic activity of peptidyltransferase. It is not clear, however, how We thank James Ebaugh for excellent technical assistance, Kathy that could account for the observed UGA specificity in Hijazi for helpful discussions, and Vishal Banthia for assistance in the suppression. Nonetheless, an intriguing observation was made construction of pFPIOO. We are particularly grateful to Warren Tate recently-namely, that deletion of two nucleotides from do- for discussions of chain termination in general and the involvement of main V, immediately adjacent to the peptidyltransferase protein LI 1 in particular, to Alexey Arkov for discussions of possible "ring," completely reverses the UGA suppression caused by interactions between 16S and 23S rRNAs, to Walter J. Pagel for A1093 (36). editorial consultation, and to Harry Noller for critical comments on In a third scenario, the A1093 mutation may increase the the manuscript. Finally, we thank M. O'Connor, A. E. Dahlberg, J. S. activity of the large ribosomal subunit protein Lii in peptide Parkinson, and J. Beckwith for strains and plasmids; Jeffrey Miller for chain termination. Experiments in vitro by Tate et al. (47) suggestions about mutD mutagenesis; and Liliana DeGeus for assis- tance in preparation of the manuscript. This research was supported demonstrated that Lii influences the activities of the two by Grant GM21499 to E.J.M. from the National Institute of General codon-specific RFs oppositely. That is, it enhances (actually, is Medical Sciences. D.K.J. was supported by a FIRST award (GM38305) essential for) the activity of RFI (specific for UAG) but from the National Institute of General Medical Sciences and by a inhibits the activity of RF2 (specific for UGA). Consequently, generous grant from the Travel Fund of Vassar College. The DNA the proposed A1093-induced increase in Lii activity would sequence analyses were performed by the M. D. Anderson Cancer lead to more inhibition of RF2-dependent termination, allow- Center's Automated DNA Sequencing Core Facility (Grant 2P30- ing more readthrough or, in effect, suppression of UGA CA16672-018 from the National Cancer Institute). mutations. Consistent with this view are the results of prelim- inary experiments to test the opposite expectation. That is, if 1. Murgola, E. J., Hijazi, K. A., Goringer, H. U. & Dahlberg, A. E. A1093 increases Lii activity, then it may increase RFi- (1988) Proc. Natl. Acad. Sci. USA 85, 4162-4165. dependent termination at UAG. Therefore, A1093 should lead 2. Noller, H. F. (1991) Annu. Rev. Biochem. 60, 191-227. to a of 3. Noller, H. F., Hoffarth, V. & Zimniak, L. (1992) Science 256, reduction suppressor tRNA-dependent readthrough of 1416-1419. UAG mutations. We observed the expected result (unpub- 4. O'Connor, M. & Dahlberg, A. E. (1993) Proc. Natl. Acad. Sci. lished observations)-namely, reduction of tRNA-dependent USA 90, 9214-9218. UAG suppression-when A1093 was introduced into strains 5. Gregory, S. T., Lieberman, K. R. & Dahlberg, A. E. (1994) containing either of two UAG-reading suppressors, one de- Nucleic Acids Res. 22, 279-284. rived from tRNALYS (lysT) and the other from tRNAGIYI (glyU). 6. Murgola, E. J. (1995) in Ribosomal RNA: Structure, Evolution, A bisulfite-induced change of G1093 to A greatly reduced the Processing and Function in Protein Synthesis, eds. Zimmermann, binding of L1I to naked mutant 23S rRNA in vitro (48). R. A. & Dahlberg, A. E. (CRC, Boca Raton, FL), pp. 359-369. Although the same investigators observed that the mutant 7. Murgola, E. J. (1985) Annu. Rev. Genet. 19, 57-80. assembles L 1I into ribosomes in vivo in a 1:1 stoichiometry, it 8. Lopilato, J., Bortner, S. & Beckwith, J. (1986) Mol. Gen. Genet. is still possible that aberrant assembly of the Ai093 50S subunit 205, 285-290. affects the functions of L1. 9. Liu, J. & Parkinson, J. S. (1989) J. Bacteriol. 171, 1254-1261. 10. Murgola, E. J. (1981) J. Mol. Biol. 149, 1-13. Finally, G1093 of 23S rRNA may be involved in an inter- 11. Buckingham, R. H., Sorensen, P., Pagel, F. T., Hijazi, K. A., action with 16S rRNA, perhaps specifically with helix 34 or nt Mims, B. H., Brechemier-Baey, D. & Murgola, E. J. (1990) 1406-1408, both of which function in chain termination (see Biochim. Biophys. Acta 1050, 259-262. refs. 6 and 34 for reviews). In support of this proposal are 23S 12. Jemiolo, D. K., Taurence, J. S. & Giese, S. (1991) Nucleic Acids rRNA crosslinks observed with 16S rRNA (49). Mutational Res. 19, 4259-4265. Downloaded by guest on September 29, 2021 Biochemistry: Jemiolo et al. Proc. Natl. Acad. Sci. USA 92 (1995) 12313 13. Gourse, R., Takebe, Y., Sharrock, R. A. & Nomura, M. (1985) 35. Moine, H. & Dahlberg, A. E. (1994) J. Mol. Biol. 243, 402-412. Proc. Natl. Acad. Sci. USA 82, 1069-1073. 36. Murgola, E. J., Pagel, F. T., Hijazi, K. A., Arkov, A. L., Xu, W. 14. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular & Zhao, S. Q. (1995) Biochem. Cell Biol. 73, in press. Cloning: A Laboratory Manual (Cold Spring Harbor Lab. Press, 37. Woese, C. R., Fox, G. E., Zablen, K., Uchida, T., Bonen, K., Plainview, NY), 2nd Ed. Pechman, K., Lewis, B. J. & Stahl, D. (1975) Nature (London) 15. Degnen, G. E. & Cox, E. C. (1974) J. Bacteriol. 117, 477-487. 254, 83-86. 16. Murgola, E. J. & Yanofsky, C. (1974) J. Bacteriol. 117, 444-448. 38. Skold, S.-E. (1983) Nucleic Acids Res. 11, 4923-4932. 17. Gutell, R. R., Schnare, M. N. & Gray, M. W. (1992) NucleicAcids 39. Schmidt, F. J., Thompson, J., Lee, K., Dijk, J. & Cundliffe, E. Res. 20, Suppl., 2095-2109. (1981) J. Bio. Chem. 256, 12301-12305. 18. Egebjerg, J., Douthwaite, S. R., Liljas, A. & Garrett, R. A. (1990) J. Mol. Biol. 213, 275-288. 40. Beauclerk, A. A. D., Cundliffe, E. & Dijk, J. (1984)J. Bio. Chem. 19. Ryan, P. C., Lu, M. & Draper, D. E. (1991) J. Mol. Biol. 221, 259, 6559-6563. 1257-1268. 41. Thompson, J., Schmidt, F. & Cundliffe, E. (1982) J. Biol. Chem. 20. Moazed, D., Robertson, J. M. & Noller, H. F. (1988) Nature 257, 7915-7917. (London) 334, 362-364. 42. Brot, N., Tate, W. P., Caskey, C. T. & Weissbach, H. (1974) Proc. 21. Moazed, D. & Noller, H. F. (1989) Cell 57, 585-597. Natl. Acad. Sci. USA 71, 89-92. 22. O'Connor, M., Goringer, H. U. & Dahlberg, A. E. (1992) Nucleic 43. Douthwaite, S., Prince, J. B. & Noller, H. F. (1985) Proc. Natl. Acids Res. 20, 4221-4227. Acad. Sci. USA 82, 8330-8334. 23. Murgola, E. J. (1995) in tRNA: Structure, Biosynthesis, and Func- 44. Douthwaite, S. (1992) J. Bacteriol. 174, 1333-1338. tion, eds. Soll, D. & RajBhandary, U. L. (Am. Soc. Microbiol., 45. Moffat, J. C., Timms, K. M., Trotman, C. N. A. & Tate, W. P. Washington, DC), pp. 491-509. (1991) Biochimie 73, 1113-1120. 24. Oeschger, M. P., Oeschger, N. S., Wiprud, G. T. & Woods, S. L. 46. Grentzmann, G., Brechemier-Baey, D., Heurgue-Hamard, V. & (1980) Mol. Gen. Genet. 177, 545-552. Buckingham, R. H. (1995) J. Biol. Chem. 270, 10595-10600. 25. Ryden, M. & Isaksson, L. A. (1984) Mol. Gen. Genet. 193,38-45. 47. Tate, W. P., Schulze, H. & Nierhaus, K. (1983)J. Bio. Chem. 258, 26. Kawakami, K., Inada, T. & Nakamura, Y. (1988)J. Bacteriol. 170, 12816-12820. 5378-5381. 48. Douthwaite, S., Vester, B., Aagaard, C. & Rosendahl, G. (1993) 27. Hughes, D., Atkins, J. F. & Thompson, S. (1987) EMBO J. 6, in The TranslationalApparatus, eds. Nierhaus, K. H., Franceschi, 4235-4239. F., Subramanian, A. R., Erdmann, V. A. & B. 28. Yano, R. & Yura, T. (1989) J. Bacteriol. 171, 1712-1717. Wittmann-Liebold, 29. Eustice, D. C., Wakem, L. P., Wilhelm, J. M. & Sherman, F. (Plenum, New York), pp. 339-346. (1986) J. Mol. Biol. 188, 207-214. 49. Mitchell, P., Osswald, M. & Brimacombe, R. (1992) Biochemistry 31, 3004-3011. 30. Dujardin, G., Pajot, P., Groudinsky, 0. & Slonimski, P. P. (1980) Mol. Gen. Genet. 179, 469-482. 50. Tate, W., Greuer, B. & Brimacombe, R. (1990) NucleicAcids Res. 31. Kruszewska, A. & Slonimski, P. P. (1984) Curr. Genet. 9, 11-19. 18, 6537-6544. 32. Fox, T. D. & Staempfli, S. (1982) Proc. Natl. Acad. Sci. USA 79, 51. Wollenzien, P., Expert-Bezancon, A. & Favre, A. (1991) Bio- 1583-1587. chemistry 30, 1788-1795. 33. Shen, Z. & Fox, T. D. (1989) Nucleic Acids Res. 17, 4535-4539. 52. Dontsova, O., Dokudovskaya, S., Kopylov, A., Bogdanov, A., 34. Brown, C. M., McCaughan, K. K. & Tate, W. P. (1993) Nucleic Rinke-Appel, J., Junke, N. & Brimacombe, R. (1992) EMBO J. Acids Res. 21, 2109-2115. 11, 3105-3116. Downloaded by guest on September 29, 2021