Purification, Cloning, and Characterization of the 16 S RNA m2G1207 Methyltransferase from Escherichia coli *

The methyltransferase that forms m2G1207 in Escherichia coli small subunit rRNA has been purified, cloned, and characterized. The gene was identified from the N-terminal sequence of the purified enzyme as the open reading frame yjjT (SWISS-PROT accession numberP39406). The gene, here renamed rsmC in view of its newly established function, codes for a 343-amino acid protein that has homologs in prokaryotes, Archaea, and possibly also in lower eukaryotes. The enzyme reacted well with 30 S subunits reconstituted from 16 S RNA transcripts and 30 S proteins but was almost inactive with the corresponding free RNA. By hybridization and protection of appropriate segments of 16 S RNA that had been extracted from 30 S subunits methylated by the enzyme, it was shown that of the three naturally occurring m2G residues, only m2G1207 was formed. Whereas close to unit stoichiometry of methylation could be achieved at 0.9 mm Mg2+, both 2 mmEDTA and 6 mm Mg2+ markedly reduced the level of methylation, suggesting that the optimal substrate may be a ribonucleoprotein particle less structured than a 30 S ribosome but more so than free RNA.

The modified nucleosides of RNA remain one of the enigmas of RNA biology. Despite being widely distributed among all classes of RNA in a bewildering plethora of shapes and forms (1), little or no functional role, and certainly no unifying concept, has yet to emerge. Although certain modulating effects on cellular processes have been described, usually under special circumstances and so far only in tRNA (2), no modified nucleoside has yet been shown to be of major importance for cell growth and/or survival. In rRNA, which next to tRNA contains the widest known assortment of modified nucleosides, no function has so far been ascribed to them, although a potential role in peptide bond formation for a subset of the modified nucleosides of Escherichia coli 23 S RNA has been postulated (3). In E. coli 16 S RNA, in which all of the modified nucleosides are known and their locations precisely determined, none was essential for any of the known ribosomal functions, although ribosomal efficiency was reduced to approximately half when all were absent (4,5). Evidence has been presented that this lowered efficiency of the 30 S subunits may be attributable to the lack of one or more of the 10 methylated residues and/or the single pseudouridine (6) in the RNA (7).
To explore the role of individual modified nucleosides of 16 S RNA in 30 S subunit structure and function, a way is needed to specifically block their formation, one at a time. The most straightforward way to do this is by inactivating the enzymes responsible for their synthesis by gene deletion or disruption, because mutation of the parent nucleoside in the RNA could have unforeseen consequences. So far, three genes and enzymes for the modified nucleosides of E. coli 16 S RNA have been described. The first, rsmA (more commonly known as ksgA), makes the two m 2 6 A residues at positions 1518 and 1519 (8). Loss of these four methyl groups has little effect on ribosome function (9). The synthase for the single pseudouridine at position 516 (6) has been deleted, 1 and the cell still grows well, although precise growth rates have not yet been determined. A methyltransferase that specifically forms m 5 C967 has recently been described, 2 but results of deletion experiments are not yet available. In this work, we describe the characterization of another methyltransferase, that for m 2 G1207. The gene was identified by N-terminal amino acid sequence analysis of the isolated enzyme and confirmed by overexpression with a His tag, affinity purification, and in vitro biochemical characterization. 3 was from Amersham Pharmacia Biotech. Plasmid pET-15b, the BL21/DE3 strain of E. coli, and His-Bind resin were from Novagen, Inc. The XL1-Blue Epicurian strain of E. coli was from Stratagene. RNase T1 was from Calbiochem, and RNase P1 was obtained from Life Technologies, Inc. DNase I was from Worthington. T4 DNA ligase and restriction enzymes were from New England Biolabs. Wizard DNA purification kits and RNasin were from Promega. BA85 cellulose nitrate filters were obtained from Schleicher & Schuell, polyvinylidene difluoride membranes were from Millipore, and omega cells were from Filtron. DEAE-Sepharose CL-6B and MonoS fast protein liquid chromatography columns were obtained from Amersham Pharmacia Biotech. Alumina type A-5 was from Sigma. Deoxyoligonucleotides for protection studies were those described previously (11). Polymerase chain reaction primers were prepared as described (12).
Purification of the m 2 G Methyltransferase-The purification procedure was similar to that previously described (13,14). One hundred grams of E. coli MRE600 cells harvested in mid-log phase were ground with 200 g of alumina. Then 3000 units of DNase I and 100 ml of Buffer A 100 containing 0.2 mM phenylmethanesulfonyl fluoride were added. The mixture was incubated for 15 min at 0°C and then centrifuged in a Sorvall SS-34 rotor for 30 min at 16,000 rpm yielding the S30. The S30 supernatant was centrifuged for 4 h at 30,000 rpm in a Beckman Ti45 rotor. The pellet was resuspended in 80 ml of Buffer A 1000 , underlayered with 40 ml of the same buffer plus 37% sucrose, and centrifuged at 45,000 rpm for 16 -18 h using slow acceleration. The supernatant of this centrifugation is the high salt wash. The high salt wash was adjusted to 35% saturation (NH 4 ) 2 SO 4 and allowed to stir on ice for 4 h. The precipitate was removed by centrifugation, and the supernatant was dialyzed overnight against Buffer B 50 . The dialyzed sample was loaded onto a 5 ϫ 43-cm DEAE-Sepharose CL-6B column equilibrated with Buffer B 50 and eluted with a linear gradient of 50 -500 mM NH 4 Cl in Buffer B 0 . The m 2 G methyltransferase activity eluted with a peak at 170 mM NH 4 Cl. The peak fractions were pooled, concentrated, and returned to Buffer B 50 by omega cell filtration. The sample was loaded onto a 0.5 ϫ 5-cm MonoS column, which had been equilibrated with Buffer B 50 . The column was eluted with a linear gradient from 50 -1000 mM NH 4 Cl in Buffer B 0 . The methyltransferase activity eluted at 250 mM NH 4 Cl. The peak fractions were pooled and either used directly or made 50% in glycerol for storage at Ϫ20°C.
Cloning and Overexpression of the m 2 G Methyltransferase Gene-The putative gene was amplified by polymerase chain reaction. The Nterminal primer extended from Ϫ13 to ϩ14, where A of the initiating ATG is ϩ1, with changes at Ϫ1, Ϫ2, and Ϫ3 to create an NdeI site adjacent to the initiating ATG. The C-terminal primer, in the reverse orientation, extended from ϩ1061 to ϩ1087, where the last sense nucleotide is 1029, and contained mismatches at 1073 and 1074 to create a BamHI site. After polymerase chain reaction, the amplified product was isolated by agarose gel electrophoresis, extracted using the Wizard DNA purification kit, concentrated by membrane filtration (Amicon Microcon 100), and digested with NdeI and BamHI. The pET-15b vector, also digested with NdeI and BamHI, and the gene insert were ligated by standard methods using T4 DNA ligase. Transformation of XL1-Blue Epicurian cells was performed by standard methods. Plasmids from three clones containing the insert were transferred into BL21/DE3 cells.
For overexpression the transformed BL21/DE3 cells were grown in Luria-Bertani broth containing 50 g/ml carbenicillin at 37°C to an A 600 of 0.7. Isopropyl-␤-D-thiogalactopyranoside at 1 mM was added to the culture, and incubation continued at 37°C to an A 600 of 1.4 -1.8. Cells were recovered by centrifugation and frozen at Ϫ70°C. For analysis of the whole cell contents, cells from 1 ml of culture were suspended in 0.03 ml of 10 mM Tris, pH 8, plus an equal volume of 2 ϫ Buffer LB, heated to 100°C for 5 min, and then chilled. For preparation of recombinant protein, cells from 100 ml of culture were washed in 10 mM Hepes, pH 7.5, and 0.17 M NaCl and sonicated in 6 ml of either 20 mM Tris, pH 8, 20 mM NH 4 Cl, 5 mM MgCl 2 , 0.1 mM EDTA, 5 mM BME, and 1 mM phenylmethanesulfonyl fluoride or Binding Buffer plus 1 mM phenylmethanesulfonyl fluoride. Both buffers gave equivalent results. The sonicated mixture was centrifuged at 15,000 ϫ g to obtain the S15 supernatant and pellet fractions. The S15 pellet was dissolved in 6 ml of Buffer D or in Binding Buffer containing 6 M urea.
Affinity Purification-The S15 supernatant was dialyzed against Binding Buffer immediately before application to a 2.5-ml column of His-Bind resin. When Binding Buffer was used in the sonication step, dialysis could be omitted. Conditions of preparation and operation of the His-Bind resin column were as described in the pET System Manual, 4th Ed. (Novagen). After addition of Elution Buffer, the tagged protein was released. The A 280 -containing fractions were pooled and dialyzed against Buffer C. The S15 pellet from the same culture was solubilized in Buffer D or Binding Buffer containing 6 M urea. When the pellet was solubilized in Buffer D, the solution was dialyzed versus Binding Buffer containing 6 M urea before being applied to a 2.5-ml His-Bind column equilibrated in the same buffer. Otherwise, it was applied directly after solubilization. Elution was with Elution Buffer plus 6 M urea. The pooled A 280 -containing fractions were dialyzed against Buffer D with decreasing concentrations of urea from 6 to 3 M at 1.0 M intervals and then from 3 to 0 M at 0.5 M intervals for 1 h each. Both protein solutions were adjusted to contain 50% glycerol and stored at Ϫ20°C.
Methylase Assays-Reaction mixtures contained 100 mM Hepes, pH 7.5, and 2 mM Mg(OAc) 2 except where otherwise indicated, 200 mM NH 4 Cl, 5 mM dithiothreitol or BME, 2 M [ 3 H]SAM, 500 units/ml RNasin, and 100 nM 16 S RNA transcript (prepared as in Ref. 15) or 80 nM 30 S ribosomal subunits reconstituted from 16 S RNA transcripts and ribosomal proteins (4,16), and enzyme as indicated. Incubation was at 37°C. The reaction was stopped by the addition of ice-cold 5% trichloroacetic acid. After precipitation at 0°C for 10 min, samples were collected on BA85 cellulose nitrate filters, the filters were dissolved in scintillation fluid, and the radioactivity was measured. One unit of activity equals 1 pmol of methyl incorporated in 30 min under the above conditions.
Nucleoside Analysis-Recombinant m 2 G methyltransferase was used to 3 H methylate 30 S particles as described above. 1.05 pmol of CH3/pmol of particles was obtained. After phenol extraction of the [ 3 H]RNA, it was digested in a 250-l incubation mixture containing 20 mM NaOAc, 0.5 mM ZnSO 4 , 40 pmol/ml carrier 16 S RNA, 80 g/ml RNase P1, and 7.3 pmol/ml [ 3 H]RNA for 2 h at 37°C. 15.4 l of 1 M Tris buffer, pH 8.0, and 5.4 units of bacterial alkaline phosphatase were added, the volume was adjusted to 275 l, and the incubation continued for an additional 2 h at 37°C. The reaction was stopped by the addition of acetic acid to a concentration of 27 mM. m 2 G was added as an internal standard, and the mixture was analyzed by HPLC (11).
Protection Studies-[ 3 H]Methyl-labeled 16 S RNA (0.36 pmol) was hybridized with a 50-fold excess of deoxyoligonucleotide in a 50-l reaction mixture, which also contained 40 mM Mes, pH 6.4, 400 mM NaCl, 9 mM EDTA, and freshly deionized 80% (v/v) formamide. The samples were heat-denatured at 90°C for 10 min and then placed at room temperature for 15 min. They were then diluted with 9 volumes of ice-cold RNase Buffer (10 mM Tris-HCl, pH 7.5, 300 mM NaCl, and 5 mM EDTA), and RNase T1 (0.2-1.8 Sankyo units) was added. The samples were digested for 30 min at 30°C. Digestion was stopped by the addition of 9 volumes of cold 10% trichloroacetic acid. After 10 min on ice, the samples were filtered on BA85 cellulose nitrate filters, the filters were dissolved, and the samples were counted.
Other Methods-N-terminal sequencing was performed on samples of the purified enzyme, which had been electrophoresed on SDS gels. The protein was electroblotted onto a polyvinylidene difluoride membrane following standard procedures (17), and N-terminal sequencing was performed as described previously (16). Protein content was assayed by a modified Bradford procedure (Bio-Rad protein assay, catalog number 500-006) using bovine serum albumin as standard. SDS gels were 12% polyacrylamide and contained 0.38 M Tris-HCl, pH 8.8, and 0.1% SDS. The 5% stacking gel contained 0.13 M Tris-HCl, pH 6.8, and 0.1% SDS. Samples were heated at 95°C for 5 min in Buffer LB and then quenched on ice for 5 min before loading. Gels were stained with Coomassie Blue.

Purification of the Methyltransferase and Identification of the
Gene-Previous work from this laboratory had identified two E. coli 16 S RNA methyltransferases, one specific for m 2 G966 and another specific for m 5 C967 (11,13,14). The m 2 G966 enzyme reacted with unmodified 30 S subunits (11,13) or unmodified 16 S RNA complexed with 30 S proteins S7 and S19 (14) but not with free 16 S RNA, whereas the m 5 C967 enzyme had the reciprocal specificity. The purpose of the present work was to identify and clone the gene for the m 2 G methyltransferase. Following a purification scheme similar, but not identical, to that previously used, a preparation that reacted with 30 S but not 16 S RNA was obtained. Gel electrophoresis showed a single strong band at ϳ37 kDa. N-terminal sequencing of this band yielded the following sequence: SAFTPASEV-LLRHSDDFEQ. A search of the GenBank data base showed an exact match to ORF yjjT (SWISS-PROT accession number P39406). The calculated molecular mass was 37.6 kDa, in good agreement with the experimental value. The presence of a strong 7-nucleotide Shine-Dalgarno sequence 6 nucleotides upstream from the initiating ATG was observed.
Cloning of the Gene, Overexpression, and Affinity Purifica-tion-The putative gene was cloned into pET-15b by standard methods. Induced cells containing clones with this insert produced large amounts of protein ϳ37 kDa in size, whereas no such band was visible in the pET control (data not shown). The overexpressed protein was distributed between the S15 supernatant fraction and S15 pellet. Distribution between these two fractions varied somewhat from preparation to preparation, and renaturation of the enzyme in the pellet produced variable results. Therefore, all experiments were conducted using the soluble supernatant fraction of the enzyme. This fraction was purified by affinity chromatography on a Ni 2ϩ -containing resin, because the overexpressed protein contained an N-terminal His tag. The eluted protein product gave a single band on gel electrophoresis at the expected size (data not shown). The finding of methyltransferase activity in the affinity-purified protein (see below) is definitive proof that the gene cloned is a methyltransferase gene. Substrate and Product Specificity of the m 2 G Methyltransferase-The overexpressed enzyme possessed the same preference for 30 S particles over free RNA as described previously (14) for the partially purified m 2 G966 methyltransferase (Fig.  1). The m 2 G enzyme was able to incorporate ϳ0.5 pmol of methyl groups/pmol of particles in 20 min, whereas 0.02 pmol of methyl groups was incorporated/pmol of free 16 S RNA in the same time interval. The nature of the methylated nucleoside was examined by HPLC. [ 3 H]methyl 30 S ribosomes were prepared using the recombinant enzyme, and the 16 S RNA was isolated. After digestion to nucleosides with RNase P1 and alkaline phosphatase, the mixture was analyzed by HPLC (Fig.  2). A single peak containing 90% of the radioactivity loaded onto the column was eluted at the m 2 G position, as confirmed by its co-elution with the added internal marker of m 2 G (Fig. 2,  arrow). These results confirm that the recombinant enzyme does catalyze formation of m 2 G in 16 S RNA.
Localization of the Site of Methylation-E. coli 16 S RNA contains three m 2 G residues, at positions 966, 1207, and 1516 (Ref. 18 and Fig. 3). To determine which G was methylated by the recombinant enzyme, hybridization-protection studies were conducted using deoxyoligomers that were complemen-tary to the RNA sequence spanning each of the m 2 G sites (13). Oligomer 958 spanned the region from residue 958 to residue 977, which included m 2 G966; oligomer 1197 spanned the region from residue 1197 to residue 1216, including m 2 G1207; and oligomer 1506 spanned the region from residue 1506 to residue 1525, which included m 2 G1516 (Fig. 3). The m 2 G methyltransferase previously characterized was specific for G966 (13). Because the purification scheme for the native enzyme used in this work was similar, and the recombinant enzyme had the same substrate specificity, we supposed that the same methyltransferase had been cloned. Unexpectedly, the hybridization-protection experiment showed clearly that this was not the case (Fig. 4). Oligomer 958 was as ineffective in protecting the label in the RNA from RNase digestion as was oligomer 1506 or no oligomer. On the other hand, oligomer 1197 efficiently protected the RNA from degradation. To control for artifacts, the deoxyoligonucleotides used in this experiment were the identical preparations used previously (13). In that work, oligomer 1197 did not protect either the m 2 G-or m 6 2 Acontaining RNAs. Therefore, its protection here cannot be attributable to some generalized inhibitory effect on RNase T1. In that work, oligomer 958 did protect both the m 5 C and m 2 G labels; thus its failure to protect here cannot be attributable to some failure of hybridization. Likewise, because oligomer 1506 did protect the m 2 6 A-containing RNA, its inactivity here cannot be attributable to a hybridization failure. We conclude from these results that the recombinant m 2 G methyltransferase described here specifically modifies G1207 rather than G966.
Mg 2ϩ Dependence of the m 2 G Methyltransferase-Because of the fact that this methyltransferase recognized 30 S ribosomes but not free RNA, we expected that Mg 2ϩ would be required for this reaction, even though a similar methyltransferase specific for m 5 C967 and for 16 S RNA functioned equally well in 1 mM EDTA as in 10 mM Mg 2ϩ . 2 Moreover, because the m 2 G966 methyltransferase studied previously was inhibited in its  (14), we tested two concentrations, 0.9 mM, which should cause partial unfolding of the particle, and 6 mM, which is known to be sufficient to stabilize natural 30 S particles in this buffer. Fig. 5 shows that at the lower Mg 2ϩ concentration, methylation approaches unit stoichiometry (0.93 mol/mol), whereas at 6 mM Mg 2ϩ , the reaction levels off at ϳ0.3 mol of methyl/mol of ribosomes. Some Mg 2ϩ is required, however, because in other experiments, 2 mM EDTA decreased the plateau level from 0.74 mol of methyl incorporated/mol of 30 S in 0.9 mM Mg 2ϩ to 0.14 mol/mol in EDTA. In these experiments, the substrate was not preincubated under the specified Mg 2ϩ conditions before the methylation reaction was initiated. In a single preincubation experiment in which the ribosomes were incubated in reaction buffer without SAM for 10 min at 37°C before initiation of the reaction with SAM, the presence of 2 or 10 mM EDTA reduced incorporation levels further to 0.1 or Ͻ0.1 mol of methyl/mol of ribosomes, respectively, whereas the control in 0.9 mM Mg 2ϩ still yielded 0.8 mol/mol.

Substrate Specificity and Mg 2ϩ
Requirements of the Enzyme-Because the m 2 G1207 methyltransferase reacts with 30 S particles but barely at all with 16 S RNA, it seems likely that methylation of the G residue occurs after the 16 S RNA has associated with some ribosomal proteins. The m 2 G966 methyltransferase, which has similar specificity, has been shown to only require the presence of proteins S7 and S19 with the 16 S RNA to be recognized by the enzyme (14). Whether the m 2 G1207 methyltransferase can also efficiently recognize 16 S RNA complexed with only a few ribosomal proteins, and what these proteins are, must await further study.
The Mg 2ϩ requirements for methylation are also consistent with the hypothesis that G1207 is methylated after the association of an unspecified number of ribosomal proteins but before assembly of the 30 S subunit is complete. At 0.9 mM Mg 2ϩ , virtually all of the G1207 residues in the 30 S subunit were available for reaction, whereas when the Mg 2ϩ concentration was raised to 6 mM, only 1 ⁄3 were still available. On the other hand, in EDTA, only 1 ⁄7 were reactive, and preincubation reduced that amount to less than 1 ⁄10. The structure of the 30 S particle at 0.9 mM Mg 2ϩ is probably partially disordered and thus may mimic the structure of the ribonucleoprotein intermediate, which may be the true substrate. In EDTA, this structure may be lost, and at 6 mM Mg 2ϩ , the 30 S particle may be too compact for reactivity. The observed partial reaction at 6 mM Mg 2ϩ may be attributable to the fact that 30 S ribosomes assembled from an RNA transcript are less stable than ones reconstituted from natural RNA (4,16). We emphasize a role for Mg 2ϩ in maintaining the proper substrate structure rather than a role at the catalytic center of the enzyme solely by analogy with the m 5 C967 methyltransferase, which had no Mg 2ϩ requirement. 2 Clearly, the current experiments do not allow us to distinguish between the two possibilities.
The substrate preference for 30 S ribosomes versus free RNA shown here for the m 2 G1207 enzyme and previously for the m 2 G966 methyltransferase (13,14) is shared by the KsgA enzyme, which also requires a 30 S subunit for reaction (20). Perhaps the ability of the m 5 C967 methyltransferase to react with free RNA, but not with 30 S subunits, is the exception to the rule, and most or all of the other 16 S RNA-modifying enzymes require either a ribonucleoprotein or a complete 30 S subunit. RsuA, which makes the single pseudouridine in 16 S RNA, also has a specific requirement for a particular RNP particle (15).
How did two similar purification procedures yield an m 2 G966 enzyme previously and an m 2 G1207 enzyme in the current work? We suspect this occurred during the DEAE column purification step. Previously the eluate was only assayed with free RNA in the belief that the m 2 G activity, measured using 30 S subunits, co-purified at that step with the m 5 C activity assayed with free RNA. However, in the present work the DEAE eluate was assayed directly with 30 S. It is possible that the m 2 G966 enzyme was overlooked in the latter case, and in the former that the m 2 G1207 enzyme did not exactly copurify with the m 5 C enzyme. Whatever the explanation, the sequenced protein identified the only ORF in E. coli with that N-terminal sequence, and this proved to be the m 2 G1207 methyltransferase.
Gene and Protein Sequence-The gene coding for the m 2 G1207 methyltransferase is yjjT. This ORF codes for a 343amino acid protein with a calculated molecular mass, 37.6 kDa, that agrees well with the value of 37 kDa for the native enzyme. We propose that the ORF be renamed rsmC for (r)ibosomal (s)mall subunit (m)ethyltransferase, because the function of the gene product has been identified as the enzyme for formation of m 2 G1207. C denotes that this is the third gene sequence for an rRNA methyltransferase to be described, the first one being ksgA, the gene for the m 2 6 A methyltransferase (8), and the second one being rsmB, the gene for the m 5 C967 methyltransferase. 2 In keeping with this nomenclature, we suggest that rsmA would be a suitable alternate name for ksgA.
The RsmC sequence contains the SAM-binding motif DXGXGXGXL (21) at residues 201-209 as well as the somewhat longer motif in which this is embedded, which has been described by Koonin et al. (22) at residues 199 -215. A search of the GenBank data base using BLAST 2.0.4 (23) identified proteins with highly significant similarity to the m 2 G1207 methyltransferase in seven organisms. These included the prokaryotes E. coli, Hemophilus influenzae, Bacillus subtilis, Staphyloccus aureus, and even the extremely divergent Thermotoga neapolitana as well as the archaebacteria Pyrococcus horikoshii and Methanococcus jannaschii. Somewhat less related were proteins from Streptomyces anulatus, Streptomyces pristinaespiralis, and Chlamydia trachomatis. No strong similarities were found with any higher eukaryotic ORFs.
The protein most similar in sequence to RsmC according to this analysis is from H. influenzae (P44453, HI0012). This protein is 330 amino acids long compared with 342 for RsmC. They are 46% identical and 65% similar along 328 residues. Given the high level of similarity between the two sequences, it is reasonable to suppose that this gene codes for an equivalent methyltransferase in H. influenzae and, therefore, that this organism also has an equivalently located m 2 G in its 16 S RNA. No information is available as to the presence of such an m 2 G in H. influenzae or for that matter in any of the other prokaryotes or archaebacteria. The next most similar amino acid sequence to RsmC is the hypothetical protein product of the E. coli ygjO gene. This 43.4-kDa protein is 33% identical and 50% similar over 173 residues to RsmC. We suspect that this protein is the m 2 G966 methyltransferase that we described previously (13,14).
Role of Methylation of G1207 in the Ribosome-m 2 G1207 is FIG. 5. Mg 2؉ dependence of methylation by the recombinant m 2 G methyltransferase. Methylation of reconstituted 30 S ribosomes was conducted as described under "Experimental Procedures," except that the incubation contained 0.9 (q) or 6 (E) mM Mg 2ϩ . The 30 S ribosome concentration was 40 nM and the purified enzyme concentration was 0.45 g/ml.  (11), and oligomer 958 -977B was the same dilution used previously (11,13). located in the helix 34 stem (see Fig. 3) in a region believed to be involved in recognition of peptide chain termination codons (24,25). This region has been directly linked to the decoding site on the ribosome by cross-links from U1052, which is adjacent to m 2 G1207 on the opposite strand to the A site codon of mRNA (26) and from A1196, 11 nucleotides away on the same strand, to the next downstream mRNA codon (10). However, no specific role for m 2 G1207 is known in chain termination or codon recognition or, for that matter, in any other function of the ribosome. Moreover, it cannot be essential for ribosome function because, as noted in the introduction, functional 30 S ribosomes have been prepared lacking all modified nucleosides. However, as also noted there, the assembly ability as well as function relative to modified controls was reduced.
To study the reasons for this reduction in assembly and function, we have embarked on a program to identify all of the genes responsible for the modifying enzymes that act on E. coli 16 S RNA. So far, we have identified three such genes, rsuA (15), rsmB, 2 and in this work, rsmC. rsmA (ksgA) was already known (8). There should be at most seven more genes to identify. Two putative genes have already been identified by their strong sequence similarity to RsmB and RsmC, respectively. Gene inactivation, one or more at a time, to block selected modifications should then help unravel the mystery of what modified nucleosides do in the ribosome.