Mutations of Non-canonical Base-pairs in the 3′ Major Domain ofEscherichia coli16 S Ribosomal RNA Affect the Initiation and Elongation of Protein Synthesis

J. Mol. Biol. (1996) 259, 207–215

Mutations of Non-canonical Base-pairs in the 3' Major Domain of Escherichia coli 16 S Ribosomal RNA Affect the Initiation and Elongation of Protein Synthesis Franc¸ois Dragon, Catherine Spickler, Robert Pinard, Julie Carrie`re and Le´a Brakier-Gingras* De´partement de Biochimie Universite´ de Montre´al Montre´al, Que´bec, H3C 3J7 Canada

Three single-point mutations and two multibase substitutions were introduced into the lower half of the 3' major domain of Escherichia coli 16 S ribosomal RNA. The three single mutations were located in helix 29 (U1341C) or in helix 43 (U1351C or A1357C) and replaced highly conserved non-canonical base-pairs with Watson-Crick base-pairs. The two multibase substitutions were located at the base of helix 42, where they transformed an irregular portion into a Waston-Crick segment. Each of the single mutations could be expressed in vivo from the rrnB operon of a multicopy plasmid under control of constitutive promoters, and none of them affected growth-rate. However, mutation A1357C, but not U1341C or U1351C, severely retarded cell growth, when expressed together with another mutation in the upper half of the 3' major domain, C1192U. The latter mutation is located in helix 34 and abolishes the binding of spectinomycin, a protein synthesis inhibitor. The proportion of mutated ribosomes was high in polysomes, suggesting that the A1357C and C1192U double mutation did not affect the initiation but rather the elongation of protein synthesis. The effect of the double mutation reveals a functional interplay between helices 34 and 43. Furthermore, an interaction between helices 34 and 43 was also suggested by studies of protection by spectinomycin against chemical attack, that showed that the binding site of spectinomycin was restored with ribosomes bearing another double mutation, U1351C and C1192U. In contrast to the single mutations, the multiple mutations in helix 42 could not be expressed in vivo under control of the strong constitutive promoters, but could be expressed under control of the weaker, thermoinducible lPL promoter. They did not affect cell growth, whether expressed in the absence or the presence of the C1192U mutation. However, under conditions where protein synthesis depended exclusively on ribosomes with plasmid-encoded rRNA, cells transformed with plasmids altered in helix 42 could not grow. Analysis of the plasmid-borne 16 S rRNA distribution in bacteria transformed with these mutant plasmids showed that mutant 16 S rRNA was present in a high proportion in the free 30 S subunits but was underrepresented in 70 S ribosomes and polysomes. Extension inhibition assays (toeprinting) demonstrated that this altered distribution resulted from an impaired capacity of the mutant 30 S subunits to form translation initiation complexes. 7 1996 Academic Press Limited

*Corresponding author

Keywords: site-directed mutagenesis; 16 S ribosomal RNA; non-canonical base-pairs; translation initiation; spectinomycin

Present address: F. Dragon, Friedrich Miescher-Institut, PO Box 2543, CH-4002 Basel, Switzerland. Abbreviations used: CAT, chloramphenicol acetyltransferase. 0022–2836/96/220207–09 $18.00/0

7 1996 Academic Press Limited

208

Introduction Ribosomal RNA is the major functional component of the ribosome, and site-directed mutagenesis of rRNA genes has been widely used to elucidate the role of rRNA in the different steps of protein synthesis (reviewed by Dahlberg, 1989; Leclerc & Brakier-Gingras, 1990; Noller, 1991; O’Connor et al., 1995). The majority of mutations that have been studied in 16 S rRNA are located in the 5', central and 3' minor domains (reviewed by Triman, 1995). A limited number of mutations have been introduced in the 3' major domain, mostly in its upper half (Makosky & Dahlberg, 1987; Go¨ringer et al., 1991; Jemiolo et al., 1991; Prescott & Kornau, 1992; Moine & Dahlberg, 1994). In this study, we investigated the effect of mutations in the lower half of the 3' major domain. The functional importance of this half-domain is well documented, and this warrants further investigation of mutations within this region. Various studies have shown that it is involved in mRNA (Bhangu & Wollenzien, 1992; Rinke-Appel et al., 1994) and tRNA binding (Moazed & Noller, 1990; Do¨ring et al., 1994; von Ahsen & Noller, 1995). A number of phylogenetically conserved non-canonical base-pairs (Gutell et al., 1994) are found within this subdomain, and most are located within the binding domain of ribosomal protein S7 (Dragon & Brakier-Gingras, 1993; see Figure 1). In this study, highly conserved non-canonical base-pairs located in helices 29, 42 and 43 (helix numbering as given by Brimacombe, 1991) were changed into Watson-Crick base-pairs by site-directed mutagenesis. The mutations were expressed in vivo in the presence or absence of an additional mutation in helix 34, which confers resistance to spectinomycin (Sigmund et al., 1984).

Results Construction of 16 S rRNA mutants Watson-Crick base-pairs were substituted for phylogenetically conserved non-canonical pairs (Noller, 1993; Gutell et al., 1994), in the lower half of the 3' major domain of Escherichia coli 16 S rRNA (Figure 1). In helix 29 (938 to 943/1340 to 1345), the universally conserved G:U pair was changed to a G:C pair (mutation U1341C). Two highly conserved non-canonical pairs in helix 43 (1350 to 1357/1365 to 1372) were changed into Watson-Crick base-pairs: a C:G pair was substituted for the U:G pair at the base of the helix (mutation U1351C), or for the A:G pair at the top of the helix (mutation A1357C). In helix 42 (1303 to 1314/1323 to 1334), four consecutive non-canonical pairs forming the base of the helix have been proposed from comparative sequence analysis (Gutell & Woese, 1990). The occurrence of this helical structure is supported by molecular probing experiments (Baudin et al., 1987). The irregular segment (1304 to 1307/1330 to 1333) was changed into a regular Watson-Crick helix by

Mutations in the 3' Major Domain of 16 S rRNA

simultaneously mutating four consecutive bases in either the 5' or the 3' strand (mutations h42/5' and h42/3', respectively).

Effect of mutations on bacterial growth The three single mutations in 16 S rRNA genes, U1341C in helix 29, U1351C and A1357C at the base and the top of helix 43, respectively, were introduced in plasmid pKK3535 or its derivatives pDVE and pSTL102. These plasmids contain the rrnB operon under control of its natural promoters P1 and P2 . The two derivatives carry a mutation at position 2058 in the 23 S rRNA (A2058G) that confers resistance to erythromycin, and plasmid pSTL102 harbors an additional mutation, C1192U, located in the upper part of helix 34 of 16 S rRNA (Figure 1), that confers resistance to spectinomycin. The presence of the mutation at position 1192 in pSTL102 enables us to selectively inactivate ribosomes containing the host-encoded 16 S rRNA by addition of spectinomycin to the culture medium. The effect of the three single mutations on growth-rate is presented in Table 1. When expressed from pKK3535, these mutations had little or no effect on the growth-rate of transformed bacteria. The same situation was observed when mutation U1341C or U1351C was introduced in pSTL102 and expressed with another mutation in 16 S rRNA, C1192U. In contrast, the simultaneous presence of the A1357C and C1192U mutations severely decreased growth rate, which suggests a functional interaction between the upper part of helix 34 and the top part of helix 43. Plasmid pSTL102 contains an additional mutation in 23 S rRNA. However, this mutation is silent and does not affect the results. Indeed, the growth of bacteria transformed with mutants derived from pDVE, which contains this 23 S rRNA mutation but not the C1192U mutation, was identical with that of bacteria transformed with mutants derived from pKK3535 (data not shown). When ribosomes containing host-encoded rRNA were inactivated by addition of spectinomycin, the pattern of growth-rate mirrored the situation observed in the absence of the drug with the double mutants U1341C plus C1192U, and A1357C plus C1192U. However, no growth was seen with the double mutant U1351C plus C1192U. Triman et al. (1989) isolated several mutants from pSTL102 that were spectinomycin-sensitive, due to the fact that the ribosomes with plasmid-encoded rRNA were non-functional. We found that the ribosomes mutated at positions 1351 and 1192 were present in large amounts in the polysomes (see below). Had they been non-functional, they would have severely affected the growth-rate in the absence of spectinomycin, which was not the case. Spectinomycin is known to protect position G1064 of 16 S rRNA against chemical attack in spectinomycin-sensitive ribosomes (Moazed & Noller, 1987). We observed that the drug also protects this position in

209

Mutations in the 3' Major Domain of 16 S rRNA

ribosomes bearing the U1351C and C1192U double mutation, showing that mutation U1351C counteracts the effect of mutation C1192U and restores spectinomycin binding (Figure 2). This accounts for the lack of growth of the double mutant in the presence of spectinomycin and, moreover, reveals an interaction between helix 34 and the base of helix 43. Two multibase substitutions were introduced at the base of helix 42. Neither mutation h42/5' nor h42/3' could be expressed from pKK3535 or pSTL102 under control of the strong constitutive promoters, suggesting that the mutations may be detrimental to the cells. They were introduced in conditional expression vectors: plasmid pLDH, which contains the rrnB operon under control of the weaker thermoinducible lPL promoter, or its derivative pLDH1192, which contains the C1192U

mutation in 16 S rRNA that confers spectinomycin resistance. Growth-rate was not significantly different from controls (41 (22) minutes) when either mutation was expressed from pLDH or pLDH1192. However, cells transformed with the mutant derivatives of pLDH1192 could not grow in the presence of spectinomycin, in which case protein synthesis depends entirely on the mutant ribosomes. Further information on this response was provided by the analysis of the distribution of the mutant plasmid-encoded 16 S rRNA in the ribosomal particles (see below). Distribution of mutant 16 S rRNA in ribosomal particles The cellular distribution of plasmid-borne 16 S rRNA was investigated by the method of Sigmund

Figure 1. Position of the single base and multibase alterations in the 3' major domain of E. coli 16 S rRNA. The secondary structure skeleton is adapted from Gutell (1993). The helices of interest are numbered according to Brimacombe (1991), and the multibranched internal loops at the base of the domain are named A and B (Dragon & BrakierGingras, 1993). The arrowhead pointing to nucleotide 1220 divides the 3' major domain into the upper part (nucleotides 986 to 1219, above the arrowhead) and the lower part (nucleotides 921 to 985 and 1220 to 1396, below the arrowhead). Substitution mutations are indicated by arrows. In helix 29, U1341 was changed to C (mutation U1341C). The irregular portion of helix 42 was changed into a Watson-Crick helix by altering its sequence in the 5' or 3' strand (mutations h42/5' and h42/3', respectively). In helix 43, U1351 was changed to C (mutation U1351C) and A1357 was changed to C (mutation A1357C). In the upper part of helix 34, the C1192 to U mutation is a genetic marker conferring resistance to spectinomycin and position G1064, which is circled, was investigated in protection studies. Positions G1338 and A1339 in loop A, to which we refer in the text, are also indicated. The broken line delineates a minimal fragment to which binds ribosomal protein S7 (Dragon & Brakier-Gingras, 1993).

210

Mutations in the 3' Major Domain of 16 S rRNA

Table 1. Growth-rate of cells transformed with plasmids coding for mutant 16 S rRNA Doubling timea (minutes) Plasmid pKK3535 pKKU1341C pKKU1351C pKKA1357C pSTL102 pSTLU1341C pSTLU1351C pSTLA1357C

Mutations in 16 S rRNA

Without spectinomycin

With spectinomycin

None U1341C U1351C A1357C C1192U C1192U + U1341C C1192U + U1351C C1192U + A1357C

35 2 3 34 2 2 39 2 2 40 2 2 33 2 2 35 2 4 40 2 5 59 2 5

No growth No growth No growth No growth 57 2 5 57 2 5 No growth 91 2 5

a Cells carrying plasmid pKK3535 or pSTL102, or their mutant derivatives, were grown at 37°C. When present, spectinomycin was added at 60 mg/ml. Values are the means 2 standard deviation of at least four experiments.

et al. (1988), after fractionating cellular lysates into 30 S, 70 S and polysomal fractions, and analyzing these fractions by primer extension and autoradiography. The relative proportion of host and plasmidborne 16 S rRNA was determined by densitometric scanning of the autoradiographs. The results are presented in Table 2. The distribution of the plasmid-encoded 16 S rRNA was similar in cells expressing plasmid pKK3535 or its mutant derivatives. The plasmidborne 16 S rRNA was incorporated to the same extent (about 70 to 80%) into the 30 S subunits, the 70 S ribosomes and the polysomes. The additional presence of mutation C1192U in pSTL102 or its mutant derivatives did not affect that distribution.

This indicates that the decrease in cell growth observed with the A1357C and C1192U double mutation does not result from a defect in subunit association or in initiation, but rather from a defect in elongation. Indeed, a defect in subunit association or in initiation would have decreased the proportion of mutant 16 S rRNA in 70 S ribosomes and in polysomes. As to the U1351C and C1192U double mutant, the proportion of mutant rRNA was high also in polysomes. We mentioned above that bacteria transformed with the plasmid having this double mutation do not grow on spectinomycin, although their growth was not affected in the absence of the drug. The high proportion of mutant rRNA in the polysomes confirms that this lack of growth cannot be ascribed to a ribosomal defect but results from the restoration of spectinomycin sensitivity. The distribution of mutant rRNA was also analyzed in cells expressing pLDH, pLDH1192 or their corresponding mutant derivatives, where the irregular portion of helix 42 was changed into a regular helix. The sucrose gradient sedimentation

Table 2. Distribution of plasmid-encoded 16 S rRNA in 30 S subunits, 70 S ribosomes and polysomes % Plasmid-encoded 16 S rRNAa Plasmidb

Figure 2. Protection of G1064 in 16 S rRNA by binding of spectinomycin to 70 S ribosomes. Ribosomes were treated with dimethylsulfate and chemical modification of guanosine residues at N-7 was detected by aniline cleavage, in the absence (lanes 1, 3 and 5) or in the presence (lanes 2, 4 and 6) of spectinomycin. Lanes 1 and 2, ribosomes isolated from bacteria transformed with plasmid pSTL102 (mutation C1192U that confers resistance to spectinomycin). Lanes 3 and 4, ribosomes isolated from bacteria transformed with pKK3535 (spectinomycinsensitive). Lanes 5 and 6, ribosomes isolated from bacteria transformed with pSTLU1351C bearing the U1351C and C1192U double mutation.

pSTL102 pSTLU1341C pSTLU1351C pSTLA1357C pLDH1192 pL1192h42/5' pL1192h42/3'

Totalc

30 S

70 S

Polysomes

82 2 5 80 2 5 78 2 7 84 2 5 49 2 6 52 2 5 48 2 5

70 2 5 75 2 5 74 2 7 76 2 5 54 2 6 65 2 5 66 2 5

76 2 5 74 2 5 69 2 5 81 2 6 53 2 6 29 2 3 32 2 4

75 2 5 74 2 5 66 2 4 81 2 6 50 2 4 29 2 5 31 2 4

a Values are the means 2 standard deviation of at least three independent preparations. b Plasmids pSTL102 and pLDH1192 and their derivatives contain the C1192U mutation that confers resistance to spectinomycin. The mutations in the derivatives of pSTL102 are described in Table 1. The derivatives of pLDH1192 contain multiple substitutions in the 5' or 3' strand at the base of helix 42. The distribution of the plasmid-encoded 16 S rRNA was similar in the absence of the C1192U mutation (data not shown). c Total cellular rRNA analyzed in crude cellular lysates.

211

Mutations in the 3' Major Domain of 16 S rRNA

profiles of ribosomal particles from cells containing mutant plasmids slightly differed from that of cells with the control plasmid, with an increase of about 40% in the pool of free subunits (data not shown). Primer extension analyses showed that the cellular distribution of plasmid-borne mutated 16 S rRNA with an altered helix 42 differs from that of the control plasmid, in that the amount of mutant plasmid-encoded 16 S rRNA was increased in free 30 S subunits and decreased in 70 S ribosomes and polysomes (Table 2). Bacteria transformed with the mutant derivatives of pLDH1192 do not grow on spectinomycin. The low proportion of mutant rRNA in the polysomes can directly account for this observation. This reduced amount of mutant ribosomes engaged in protein synthesis is unable to support bacterial growth when protein synthesis depends exclusively on them. The increased proportion of free 30 S subunits with mutant rRNA demonstrates that perturbing the irregular portion of helix 42 lowers the capacity of the 30 S subunits to participate in protein synthesis, and suggests an association and/or initiation defect. Functional assays The 30 S subunits mutated in helix 42 were further investigated for their ability to associate to 50 S subunits and to bind a messenger RNA. The rate of formation of the 30 S initiation complex was also assessed, using the extension inhibition assays (toeprinting), developed by Hartz et al. (1988). In these assays, a primer annealed downstream from the initiation codon of a messenger RNA is extended with reverse transcriptase. The enzyme stalls when it encounters the 30 S initiation complex. The relative toeprint, which is the ratio of the amount of truncated transcript to that of the truncated plus full-length transcripts reflects the rate of formation of the ternary complex between 30 S subunits, mRNA and the initiator tRNA (Spedding et al., 1993). No difference was seen between mutant and wild-type 30 S subunits when assessing the association with 50 S subunits and the binding of a messenger RNA (data not shown). However, as shown in Figure 3 and Table 3, we found that with

Figure 3. Example of toeprinting assays with 30 S subunits and CAT mRNA. Lane 0, control lacking 30 S subunits; lane 1: 30 S subunits mutated in helix 42, isolated from bacteria transformed with pL1192h42/5'; lane 2, control 30 S subunits isolated from bacteria transformed with pLDH1192. Similar results were obtained with 30 S subunits originating from cells transformed with pL1192h42/3' (not shown). The presence or the absence of the C1192U mutation did not affect the toeprints (not shown). The arrowheads indicate the toeprint signal corresponding to the inhibition of extension (the lower band) and the full-length transcript (the topmost band).

30 S subunits mutated in helix 42, the relative toeprint was decreased by about 40% when compared with wild-type 30 S subunits. This represents a drastic effect considering that the experiments were performed with a mixed population in which about 65% of the 30 S subunits contained the mutated 16 S rRNA. This observation demonstrates that the irregular portion of helix 42 is critical for initiation. The presence or the absence of the C1192U mutation did not affect the results.

Table 3. Formation of the 30 S initiation complex with 30 S subunits mutated in helix 42 of 16 S rRNA Origin of the 30 S subunitsa Cells transformed with: pLDH1192 (control) pL1192h42/5' pL1192h42/3'

Mutations in 16 S rRNA

Relative toeprintsb

C1192U C1192U + 5' side of helix 42 C1192U + 3' side of helix 42

0.84 2 0.05 0.55 2 0.06 0.49 2 0.07

a The proportion of plasmid-encoded 16 S rRNA was about 65% in 30 S subunits isolated from cells transformed with the mutant plasmid (see Table 2). b The relative toeprints correspond to the intensity of the truncated transcript band (toeprint signal) divided by the sum of the intensities of the truncated and full-length transcript bands measured by densitometric scanning of the autoradiographs. Similar results were obtained with wild-type and mutant 30 S subunits without the C1192U mutation. Results are the means 2 standard deviation of six independent assays.

212

Discussion Replacement of single non-canonical base-pairs with Watson-Crick pairs in the lower part of the 3' major domain of E. coli 16 S rRNA Three single-point mutations in helices 29 and 43 transformed non-canonical base-pairs into WatsonCrick pairs in the 3' major domain of E. coli 16 S rRNA. These mutations could be expressed constitutively and had little effect on growth-rate. This contrasts with the view that most G:U pairs and other non-canonical pairings in 16 S rRNA are not accidental events and could play a role in ribosome function (Gutell et al., 1994). The observation that the replacement of a universally conserved base does not affect ribosome function is not unprecedented, as shown, for example, when replacing C1400 with U or U531 with G in 16 S rRNA (Thomas et al., 1988; Santer et al., 1993). However, to our knowledge, this is the first reported study of mutations that replace non-canonical base-pairs with Watson-Crick base-pairs. Our results suggest that the geometry of the helices where the substitutions were introduced is not significantly perturbed. Our studies with these mutations were performed under optimal conditions of growth, in a rich medium and at 37°C. We cannot exclude that the mutations could affect cell growth under different conditions, such as a poor medium, or either lower or higher temperatures. Two of the mutations, U1341C and U1351C, were previously shown to respectively increase or decrease about twofold the attachment of S7 to a minimal fragment encompassing its binding site within the lower part of the 3' major domain (Dragon et al., 1994). It is likely that in vivo, the cooperative ribosome assembly overcomes these effects, since they are not mirrored by significant changes in growth-rate. A similar observation was made by Rosendahl & Douthwaite (1995), when comparing the effects of mutations in 23 S rRNA on the binding of L11 in vitro and in vivo. Mutation A1357C, which substitutes a C:G pair for the A:G pair closing the terminal loop of helix 43, affects elongation of protein synthesis when combined with mutation C1192U in helix 34 that causes resistance to spectinomycin. It has been deduced from protection and mutagenesis studies that spectinomycin binds to the upper part of helix 34, in the vicinity of the G1064-C1192 base-pair (Moazed & Noller, 1987; Samaha et al., 1994; Johanson & Hughes, 1995 and references therein). The antibiotic blocks translocation by inhibiting the effect of elongation factor EF-G (Bilgin et al., 1990), which promotes the displacement of the anticodon end of the tRNA along with the mRNA from the A to the P site in the 30 S subunit (Moazed & Noller, 1989). The folding of the 16 S rRNA transiently brings the region around position 1050 in the lower part of helix 34 into the vicinity of the decoding site

Mutations in the 3' Major Domain of 16 S rRNA

during protein synthesis (Donstova et al., 1992), and this localization could account for the effect of spectinomycin on the movement of the anticodon end of tRNA. A mutation at position 1192 alone does not affect translocation but it is tempting, although speculative at this point, to suggest that the A1357C and C1192U double mutation affects translocation, implying that an interplay between these two sites has a role in this step. The terminal loop of helix 43 interacts with the mRNA a few nucleotides upstream from the P-site codon (Rinke-Appel et al., 1994), and mutation A1357C probably affects the conformation of the loop. How position 1192 in the upper part of helix 34 interacts with the base-pair closing the loop of helix 43 remains to be determined. A ribosomal protein could be involved in this interaction. The U1351C mutation at the base of helix 43 combined with the C1192U mutation did not affect the activity of the ribosomes but restored spectinomycin sensitivity. Ribosomes with the double mutation exhibit the same pattern of protection by spectinomycin against chemical attack as wild-type ribosomes. This demonstrates that the mutated ribosomes recovered the initial binding site for the drug. If the double mutation had created a novel binding site for spectinomycin, a different pattern of protection would have been observed. We therefore conclude that perturbing or changing the noncanonical G:U pair at the base of helix 43 affects helix 34, which restores the structure required for spectinomycin binding.

Replacement of multiple non-canonical base-pairs with Watson-Crick pairs in the lower part of the 3' major domain of E. coli 16 S rRNA The two multibase substitutions h42/5' and h42/3' were designed to convert the irregular helical segment at the base of helix 42 into a Watson-Crick helix. Neither mutation could be expressed constitutively with the natural promoters of the rrnB operon. When induced with a weaker promoter, they both conferred a deleterious phenotype, when the burden of protein synthesis relied exclusively on the mutated ribosomes. Extension inhibition assays indicated that the capacity of 30 S subunits mutated in helix 42 to form 30 S initiation complexes was severely impaired. Since the binding of messenger RNA to the 30 S subunits was not affected, we can propose that a decreased binding of the initiator tRNA at the P site impairs the formation of the 30 S initiation complex. Indeed, several studies based on protection, modification interference and crosslinking experiments have shown that G1338 and A1339 in loop A, in proximity to the irregular portion of helix 42 (Figure 1), are involved in tRNA binding at the P site (Moazed & Noller, 1990; Do¨ring et al., 1994; von Ahsen & Noller, 1995). Interestingly, when we studied the minimal fragment within the 3' major

213

Mutations in the 3' Major Domain of 16 S rRNA

domain that interacts with S7, we observed that deleting the irregular portion of helix 42 abolishes the reactivity of G1338 and A1339. However, when the irregular structure of helix 42 was turned into a regular Watson-Crick helix, the reactivity of these two positions was not significantly affected (Dragon et al., 1994). The conditions of high ionic strength that were used for chemical probing probably prevent changes that occur under more physiological conditions. Therefore, our interpretation for the effect of mutations h42/5' and h42/3' is that the irregular portion of helix 42 modulates the conformation of the P site in the 30 S subunits. We conclude that mutations perturbing this irregular portion interfere with the binding of tRNA at the P site by altering loop A, thus restricting the access of the mutated 30 S subunits to the pool of active polysomes.

Materials and Methods Bacterial strains and plasmids Plasmids pDVE (Vester & Garrett, 1987) and pSTL102 (Triman et al., 1989), which contain the entire rrnB operon from E. coli under control of its natural constitutive promoters, are two derivatives of pKK3535 (Brosius et al., 1981). Both pDVE and pSTL102 carry a mutation in the 23 S rRNA gene, an A to G transition at position 2058 (A2058G) that confers resistance to erythromycin. Plasmid pSTL102 contains an additional mutation in the 16 S rRNA gene, a C to U transition at position 1192 of 16 S rRNA (C1192U) that confers resistance to spectinomycin. Plasmids pLDH and pLDH1192 are gifts from Drs A. E. Dahlberg and M. O’Connor (Brown University, Providence, RI). They are derived from pNO2680 (Gourse et al., 1985) which contains the rrnB operon under control of the thermoinducible lPL promoter. Plasmid pLDH1192 differs from pLDH by the presence of the C1192U marker in 16 S rRNA. E. coli strain MC140, F− D(lac-pro), thi−, recA− (Moine & Dahlberg, 1994), carrying the temperature-sensitive repressor cI857 on a neomycin/kanamycinresistant plasmid, pLG857 (O’Connor et al., 1992), was used as a host for all plasmids containing the rrnB operon under control of the lPL promoter, and E.coli strain DH1 (Hanahan, 1983) was used as a host for all plasmids containing the rrnB operon under control of its constitutive promoters.

Growth-rate measurements Bacterial cultures were grown in 125 ml flasks containing 25 ml of LB medium supplemented with glucose at 0.1% (w/v) and ampicillin (Boehringer Mannheim) at 200 mg/ml. Cells transformed with plasmids containing the rrnB operon with the constitutive promoters were grown at 37°C, and cells transformed with plasmids containing the lPL promoter were grown overnight at 30°C, then at 42°C to late log phase. When required, LB medium was supplemented with spectinomycin (Sigma) at 60 mg/ml. Bacterial growth-rates were determined as described by Miller (1972), by following the absorbance at 600 nm. Analysis of mutant 16 S rRNA distribution Ribosomal particles were isolated from cellular lysates prepared by freezing and thawing, essentially as described (Ron et al., 1966). Lysates (15 A260 units) were applied onto 10% to 40% (w/v) linear sucrose gradients, and centrifuged at 35,000 rpm for three hours in a Beckman SW41 rotor at 4°C. Fractions containing 30 S subunits, 70 S ribosomes and polysomes were collected using an ISCO gradient fractionator. The rRNA was isolated from the fractions by extraction with phenol, and the proportion of plasmid and host-encoded 16 S rRNA was determined by primer extension (Sigmund et al., 1988), using either a primer specific for the spectinomycin resistance mutation C1192U or for the mutations introduced by site-directed mutagenesis. The extent of plasmid-encoded rRNA in total cellular rRNA was determined in crude cellular lysates. The relative band intensities from lightly exposed X-ray films (Fuji RX) were measured with an LKB UltroScan XL laser densitometer. Assessment of spectinomycin binding to 70 S ribosomes by chemical modification Ribosomes (100 pmol) were incubated in 100 ml of binding buffer (50 mM Hepes-KOH (pH 7.8), 20 mM MgCl2 , 100 mM NH4 Cl, 1 mM dithiothreitol, 0.5 mM EDTA) for 30 minutes at 37°C in the presence or absence of 20 molar equivalents of spectinomycin, and then for ten minutes on ice. Modification of guanosine residues at N-7 was detected after treatment with dimethylsulfate and aniline-induced strand scission, as described by Moazed & Noller (1987). After stopping the reaction, the rRNA was extracted and analysed by primer extension and electrophoresis on 8% (w/v) polyacrylamide sequencing gels.

Site-directed mutagenesis

Functional assays

Point mutations and multibase substitutions were introduced by site-directed mutagenesis (Kunkel et al., 1987) in a pBluescript (Stratagene) intermediate, that contains a KpnI-XbaI fragment from pNO2680 encoding the 16 S rRNA gene, with or without the C1192U mutation. Each mutant phagemid was characterized by dideoxy sequencing (Sanger et al., 1977). Mutant ApaI-XbaI fragments, with or without the C1192U marker, were subcloned into the expression vectors pSTL102 and pLDH1192, or pKK3535, pDVE and pLDH, respectively. The structure of all plasmid derivatives was verified by restriction enzyme analysis and sequencing of the mutagenized portion.

Mutated 30 S subunits with a decreased capacity to participate in protein synthesis were investigated in various functional assays. Their capacity to associate to 50 S subunits was assessed by incubating 1.0 A260 unit of 30 S subunits with 2.0 A260 units of 50 S subunits in 0.2 ml of the association buffer (50 mM Hepes-KOH (pH 7.5), 50 mM NH4 Cl, 6 mM magnesium acetate, 5 mM dithiothreitol) for 30 minutes at 37°C. The formation of 70 S ribosomes was analyzed by centrifugation for 16 hours in a SW41 rotor at 35,000 rpm through a 15% to 40% sucrose gradient made in the same buffer . The binding of a messenger RNA to the 30 S subunits was assessed by filtration on nitrocellulose membranes with the messen-

214 ger coding for chloramphenicol acetyltransferase (CAT). The CAT messenger was generated by in vitro transcription from plasmid pLRCAT, as described by Pinard et al. (1995). The messenger was labeled with [a-35S]UTP, and binding assays were performed as described (Melanc on et al., 1990). Extension inhibition assays (toeprinting) were used to measure the rate of formation of the 30 S initiation complex (Hartz et al., 1988) with the CAT messenger RNA. The 30 S subunits used in the assays were obtained by fractionation of cell lysates, as described above. Extension inhibition was performed with a primer complementary to positions 76 to 96 downstream from the initiation codon of the CAT mRNA. Toeprinting results were quantified by measuring band intensities of the full-length and truncated (toeprint signal) transcripts with an LKB UltroScan XL laser densitometer.

Acknowledgements We are grateful to Robert Cedergren, Michael O’Connor, Michael Laughrea, Kathleen Triman and Louise Wickham for stimulating comments, and to Stephen Douthwaite for the restriction map and computerized sequence of pSTL102. F.D. and R.P. held a studentship from the FCAR (Fonds pour la Formation de Chercheurs et l’Aide a` la Recherche). This work was supported by the Medical Research Council of Canada.

References Baudin, F., Ehresmann, C., Romby, P., Mougel, M., Colin, J., Lempereur, L., Bachellerie, J.-P., Ebel, J.-P. & Ehresmann, B. (1987). Higher-order structure of domain III in Escherichia coli 16 S ribosomal RNA, 30 S subunit and 70 S ribosome. Biochimie, 69, 1081–1096. Bhangu, R. & Wollenzien, P. (1992). The mRNA binding track in the Escherichia coli ribosome for mRNAs of different sequences. Biochemistry, 31, 5937–5944. Bilgin, N., Richter, A. A., Ehrenberg, M., Dahlberg, A. E. & Kurland, C. G. (1990). Ribosomal RNA and protein mutants resistant to spectinomycin. EMBO J. 9, 735–739. Brimacombe, R. (1991). RNA-protein interactions in the Escherichia coli ribosome. Biochimie, 73, 927–936. Brosius, J., Ullrich, A., Raker, M. A., Gray, A., Dull, T. J., Gutell, R. R. & Noller, H. F. (1981). Construction and fine mapping of recombinant plasmids containing the rrnB ribosomal operon of E. coli. Plasmid, 6, 112–118. Dahlberg, A. E. (1989). The functional role of ribosomal RNA in protein synthesis. Cell, 57, 525–529. Donstova, O., Dokudovskaya, S., Kopylov, A., Bogdanov, A., Rinke-Appel, J., Ju¨nke, N. & Brimacombe, R. (1992). Three widely separated positions in the 16 S RNA lie in or close to the ribosomal decoding region; a site-directed cross-linking study with mRNA analogues. EMBO J. 11, 3105–3116. Do¨ring, T., Mitchell, P., Osswald, M., Bochkariov, D. & Brimacombe R. (1994). The decoding region of 16 S RNA; a cross-linking study of the ribosomal A, P and E sites using tRNA derivatized at position 32 in the anticodon loop. EMBO J. 13, 2677–2685. Dragon, F. & Brakier-Gingras, L. (1993). Interaction of Escherichia coli ribosomal protein S7 with 16 S rRNA. Nucl. Acids Res. 21, 1199–1203. Dragon, F., Payant, C. & Brakier-Gingras L. (1994). Mutational and structural analysis of the RNA

Mutations in the 3' Major Domain of 16 S rRNA

binding site for Escherichia coli ribosomal protein S7. J. Mol. Biol. 244, 74–85. Go¨ringer, H. U., Hijazi, K. A., Murgola, E. J. & Dahlberg, A. E. (1991). Mutations in 16 S rRNA that affect UGA (stop codon)-directed translation termination. Proc. Natl Acad. Sci. USA, 88, 6603–6607. Gourse, R. L., Takebe, Y., Sharrock, R. & Nomura, M. (1985). Feedback regulation of rRNA and tRNA synthesis and accumulation of free ribosomes after conditional expression of rRNA genes. Proc. Natl Acad. Sci. USA, 82, 1069–1073. Gutell, R. R. (1993). Collection of small subunit (16 S- and 16 S-like) ribosomal RNA structures. Nucl. Acids Res. 21, 3051–3054. Gutell, R. R. & Woese, C. R. (1990). Higher order structural elements in ribosomal RNAs: pseudoknots and the use of noncanonical pairs. Proc. Natl Acad. Sci. USA, 87, 663–667. Gutell, R. R., Larsen, N. & Woese, C. R. (1994). Lessons from an evolving rRNA: 16 S and 23 S rRNA structures from a comparative perspective. Microbiol. Rev. 58, 10–26. Hanahan, D. (1983). Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166, 557–580. Hartz, D., McPheeters, D. D., Traut, R. & Gold, L. (1988). Extension inhibition analysis of translation initiation complexes. Methods Enzymol. 164, 419–425. Jemiolo, D. K., Taurence, J. S. & Giese, S. (1991). Mutations in 16 S rRNA in Escherichia coli at methyl-modified sites: G966, C967, G1207. Nucl. Acids Res. 19, 4259–4265. Johanson, U. & Hughes, D. (1995). A new mutation in 16 S rRNA of Escherichia coli conferring spectinomycin resistance. Nucl. Acids Res. 23, 464–466. Kunkel, T. A., Roberts, J. D. & Zakour, R. A. (1987). Rapid and efficient site-specific mutagenesis without phenotypic selection. Methods Enzymol. 154, 367–382. Leclerc, D. & Brakier-Gingras, L. (1990). Study of the function of Escherichia coli ribosomal RNA through site-directed mutagenesis. Biochem. Cell Biol. 68, 169–179. Makosky, P. C. & Dahlberg, A. E. (1987). Spectinomycin resistance at site 1192 in 16 S ribosomal RNA of E. coli: an analysis of three mutants. Biochimie, 69, 885–889. Melanc on, P., Leclerc, D. & Brakier-Gingras, L. (1990). A deletion at the 5' end of Escherichia coli 16 S ribosomal RNA. Biochim. Biophys. Acta, 1050, 98–103. Miller, J. H. (1972). Determination of viable cell counts: bacterial growth curves. In Experiments in Molecular Genetics, pp. 31–36, Cold Spring Harbor Laboratory Press, Plainview, NY. Moazed. D. & Noller, H. F. (1987). Interaction of antibiotics with functional sites in 16 S ribosomal RNA. Nature, 327, 389–394. Moazed, D. & Noller, H. F. (1989). Intermediate states in the movement of transfer RNA in the ribosome. Nature, 342, 142–148. Moazed, D. & Noller, H. F. (1990). Binding of tRNA to the ribosomal A and P sites protects two distinct sets of nucleotides in 16 S rRNA. J. Mol. Biol. 211, 135–145. Moine, H. & Dahlberg, A. E. (1994). Mutations in helix 34 of Escherichia coli 16 S ribosomal RNA have multiple effects on ribosome function and synthesis. J. Mol. Biol. 243, 402–412. Noller, H. F. (1991). Ribosomal RNA and translation. Annu. Rev. Biochem. 60, 191–227.

Mutations in the 3' Major Domain of 16 S rRNA

Noller, H. F. (1993). On the origin of the ribosome: coevolution of subdomains of tRNA and rRNA. In The RNA World (Gesteland, R. F. & Atkins, J. F., eds), pp. 137–156, Cold Spring Harbor Laboratory Press, Plainview, NY. O’Connor, M., Go¨ringer, H. U. & Dahlberg, A. E. (1992). A ribosomal ambiguity mutation in the 530 loop of E. coli 16 S rRNA. Nucl. Acids Res. 20, 4221–4227. O’Connor, M., Brunelli, C. A., Firpo, M. A., Gregory, S. T., Lieberman, K. R., Lodmell, J. S., Moine, H., Van Ryk, D. I. & Dahlberg, A. E. (1995). Genetic probes of ribosomal RNA function. Biochem. Cell Biol. 73, 859–868. Pinard, R., Payant, C. & Brakier-Gingras, L. (1995). Mutations at positions 13 and/or 914 in Escherichia coli 16 S ribosomal RNA interfere with the initiation of protein synthesis. Biochemistry, 34, 9611–9616. Prescott, C. D. & Kornau, H.-C. (1992). Mutations in E. coli 16 S rRNA that enhance and decrease the activity of a suppressor tRNA. Nucl. Acids Res. 20, 1567–1571. Rinke-Appel, J., Ju¨nke, N., Brimacombe, R., Lavrik, I., Dokudovskaya, S., Donstova, O. & Bogdanov, A. (1994). Contacts between 16 S ribosomal RNA and mRNA, within the spacer region separating the AUG initiator codon and the Shine-Dalgarno sequence; a site-directed cross-linking study. Nucl. Acids Res. 22, 3018–3025. Ron, E. Z., Kohler, R. E. & Davis, B. D. (1966). Polysomes extracted from Escherichia coli by freeze-thawlysozyme lysis. Science, 153, 1119–1120. Rosendahl, G. & Douthwaite, S. (1995). Cooperative assembly of proteins in the ribosomal GTPase centre demonstrated by their interactions with mutant 23 S rRNAs. Nucl. Acids Res. 23, 2396–2403. Samaha, R. R., O’Brien, B., O’Brien, T. W. & Noller, H. F. (1994). Independent in vitro assembly of a ribonucleoprotein particle containing the 3' domain of 16 S rRNA. Proc. Natl Acad. Sci. USA, 91, 7884–7888.

215 Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. Natl Acad. Sci. USA, 74, 5463–5467. Santer, U. V., Cekleniak, J. A. & Santer, M. (1993). The 530 loop of 16 S ribosomal RNA. Each base change confers its own phenotype. Mol. Biol. Cell, 4, 420a. Sigmund, C. D., Ettayebi, M. & Morgan, E. A. (1984). Antibiotic resistance mutations in 16 S and 23 S ribosomal RNA genes of E. coli. Nucl. Acids Res. 12, 4653–4663. Sigmund, C. D., Ettayebi, M., Borden, A. & Morgan, E. A. (1988). Antibiotic resistance mutations in ribosomal RNA genes of Escherichia coli. Methods Enzymol. 164, 673–690. Spedding, G., Gluick, T. C. & Draper, D. E. (1993). Ribosome initiation complex formation with the pseudoknotted a operon messenger RNA. J. Mol. Biol. 229, 609–622. Thomas, C. L., Gregory, R. J., Winslow, G., Muto, A. & Zimmermann, R. A. (1988). Mutations within the decoding site of Escherichia coli 16 S rRNA: growthrate impairment, lethality and intragenic suppression. Nucl. Acids Res. 16, 8129–8146. Triman, K. L. (1995). Mutational analysis of 16 S ribosomal RNA structure and function in Escherichia coli. Advan. Genet. 33, 1–39. Triman, K., Becker, E., Dammel, C., Katz, J., Mori, H., Douthwaite, S., Yapijakis, C., Yoast, S. & Noller, H. F. (1989). Isolation of temperature-sensitive mutants of 16 S rRNA in Escherichia coli. J. Mol. Biol. 209, 645–653. Vester, B. & Garrett, R. A. (1987). A plasmid-coded and site-directed mutation in Escherichia coli 23 S RNA that confers resistance to erythromycin: implications for the mechanism of action of erythromycin. Biochimie, 69, 891–900. von Ahsen, U. & Noller, H. F. (1995). Identification of bases in 16 S rRNA essential for tRNA binding at the 30 S ribosomal P site. Science, 267, 234–237.

Edited by M. Gottesman (Received 27 December 1995; received in revised form 23 February 1996; accepted 18 March 1996)