Effects of mutagenesis of a conserved base-paired site near the decoding region of Escherichia coli 16 S ribosomal RNA

J. Mol.

Biol.

(1990) 212, 127-133 --

ffects of Mutagenesis of a Conserved Base-paired Near the Decoding Region of Escherichia coli 16 S Ribosomal RNA E. A. De Stasio”f and A. E. Dahlberg Division

of Biology and Medicine Brown University Providence, RI 02912, U.S.A.

(Received 3 July

1989; accepted Novernbey 1989)

Eleven of 15 possible single and double mutations were constructed in a cloned copy of Exherichia coli 16 S rDNA at a base-paired site, 1409-1491. Expression of any of these mutations was detrimental to the growth of E. co&. Mutations that substituted unpaired purine bases were lethal in the system described. Otherwise, the degree of detrimental effect on growth-rate was not directly correlated with specific rRNA primary or secondaIry structures. Using reverse transcription of rRNA isolated from subunits or 70 S ribosomes, we wer’e able to determine the amount of mutant rRNA used in translation. From these experiments, we found that the lethal mutations appeared to be selectively excluded from the pool of 70 S ribosomes following expression from a repressible plasmid. In contrast, a non-lethal mutation was present in subunits, ribosomes and polysomes in approximately equal amounts. Mutations that disrupted base-pairing were found to confer varying levels of resistance to nine aminoglycosides, including four neomycins, two kanamycins, gentamicin, apramycin and hygromycin. A high frequency of reversion from resistant and slow-growth phenotypes due to a host mutation was observed,

1. Introduction

number of other highly conserved nucleotides. Antibiotics that interfere with decoding fun&ions also protect nucleotides in this same region from chemical modification (Moazed & Noller, 1987). Mutations and modifications of rRNA that confer resistance to these antibiotics also reside in the decoding region (for a review, see De Stasio et al., 1988). The decoding region of 16 S rRNA has been studied by several groups using the approa,ch of sitedirected mutagenesis. Thomas et al. (1988) have produced a number of C to U transition mutations in the 1400 region that yield a variety of phenotypes when expressed in vivo. Rottmann et al. (1988) reported that 16 S rRNA containing a double mutation (1399 and 1401) was lethal when expressed from natural rrnB promoters, while a viable G to U mutation at 1416 was simply underrepresented in the 70 S fraction. Hui et ial. (1988) reported that mutations of Cl400 deereaw translation of human growth hormone mRNA in their specialized ribosome system, while Denman et al. (1989) found effects of mutations around 1400 in their in vitro reconstituted system only when nucleotides were inserted or deleted. While none of these reports conclusively defines the functional role

Recent studies have demonstrated the importance of rRNA for ribosome function. Specific sequences of 16 S ribosomal RNA have been shown to be essential for proper translation initiation (Hui & de Boer, 1987; Jacob et al., 1987) and for accurate chain termination (Murgola et al., 1988). A growing number of other regions in both 16 S and 23 S rRNA have been identified with specific functions (for reviews, see De Stasio et al., 1988; Dahlberg, 1989). One area of particular interest is the putative decoding site in the 30 S subunit. This consists of several conserved regions of 16 S rRNA, held together in a complex, higher-ordered structure, which together comprise the site of tRNA-mRNA interaction. Moazed & Noller (1986) have identified the precise bases protected by tRNA bound to the ribosome by chemical footprinting. They include the single-stranded regions 1397 to 1408 and 1492 to 1505 of I6 S rRNA (see Fig. 1) in addition to a

7 Present address: Department of Genetics, 445 Henry Mail, University of Wisconsin-Madison, Madison, WI 53706, 7T.S.A. 127 0022-2836~90~020127-07

$03.00/O

0 1990 Academjc Press Limited

E. A. De Btasio and A. E. Dahlberg

128

provided by Eli Lilly; neamine was from Upjohn; hygromycin was a kind gift from E. Cundliffe.

1400

(b) Bacterial

- c G i Cm C C G u

G

1409

1418-

A - C-G A-U C-G C-G A 0 U-G GmU G-C GoA A 0 I

u

A

kc A .

,491

G

c G

PllU A A - 1500 c A A G G

G-1402 I

Figure 1. Sites of mutagenesis, 140991491, in E. coli 16 S rRNA. The outline of 16 S rRNA (inset) is from Stern et al. (1988).

of specific bases in the 1400 region, it is clear that certain sequences are important for proper ribosome function. The present study focusses on one “end” of the decoding site, the base-pair (C1409.Gl491) that links the two highly conserved single-stranded regions of the decoding site (Fig. 1). Nucleotides 1409 and 1491 reside at the base of the irregular penultimate helix near the 3’ end of 16 S r-RNA. The evolutionary conservation of this pair is not absolute; several eukaryotes have an U. A pair and halobacteria have an A. U pair at the equivalent position. In addition, Aspergillus mitochondrial rRNA has a G.U pair at this site (Gutell et aZ., 1985). Although conservation is not universal, one might expect that changes in secondary structure at the base of the helix (at one end of the decoding site) may have severe effects on ribosome function. Eleven of the 15 possible mutant nucleotide juxtaat base-pair 1409-1491 have been positions constructed. To our surprise, the 11 mutations yield a variety of phenotypes when expressed in Escherichia coli. One phenotype, paromomycin resistance conferred by certain of these mutations, has been described (De Stasio et al., 1989).

2. Materials (a) Enzymes

and Methods and reagents

Restriction enzymes, bacteriophage T4 DNA ligase and Klenow fragment were purchased from New England Biolabs. Radiochemicals were from NEN and nucleotides were from BRL. Uracil n-glycosylase was a generous gift from B. Duncan. Paromomycin and apramycin were

strains

and

and

plasmids

Strain XL-1 Blue (Stratagene) was used for work with M13. Strain RZ1032 (ungg, dutt) was a gift from T. Kunkel. Ribosomal mutants were propagated in strains HBlOl (Bolivar & Bachman, 1979) and DHI (Ranahan, 1983). Plasmid pKK3535, containing the entire rrnB operon, has been described (Brosius et al., 1981) as has pN02680, in which PI and P2 have been replaced by PL from phage lambda (Gourse et aE., 1985). Both plasmids marker. The plasmid carry an ampicillin-resistance pCI857 (Remaut et aE., 1981) carries a gene coding for the temperature-sensitive repressor ~1857 and a neomycinresistance marker. MISmpl9 was purchased from BRL. (c) Mutagenesis Site-directed mutagenesis was performed using 17-base oligonucleotides complementary to the minus strand of the 16 S rDNA gene. All 3 mutagenic bases were incorporated at the mutant position. Mutagenesis was performed using the l-primer method of Zoller & Smith (1984) with the modifications of Kunkel (1985). Singlestranded template Ml3 DNA, containing the EcoRIIXbaI fragment of I6 S rDNA was prepared from strain RZ1032 to allow the incorporation of uridine. Following primer extension and ligation, the DNA was treated with 10 ng of uraeil n-glycosylase for 30 min at 37 “C! to create abasic sites. The non-mutagenic strand was further destroyed by sodium hydroxide addition to pH 12%. After buffering to pH 8.0, reactions were used to transform strain XL-I. Screening for mutations was done solely by sequencing according to Sanger et al. (1977). The frequency of mutagenesis was approximately 80 to 90%. Double mutants were produced by using the NcoI site between 1409 and 1491 and cloning the NcoI-XbaI fragment containing 1491 mutations into appropriate Ml3 vectors containing 1409 mutations. Cloning into expression vectors was done using the unique AyaI and XbaI sites in 16 S rDNA. To be sure the mutations had been cloned, plasmid DNA was prepared and the EcoRIIXbaI fragment of 16 S rDNA was cloned back to Ml3 for sequencing. (d) Bacterial growth and aminoglycoside resistatzce transformed with mutant plasmids were in Luria-Bertani medium supplemented with 200 pg ampicillin/ml (and 50 pg neomycin/ml if selection for pCI857 was required). Growth rate and antibiotic resistance experiments were always begun with cells from an agar plate from which fast-growing revertants were removed. Minimal inhibitory concentrations were determined by the tube dilution method, using 2-fold serial dilutions of antibiotics ranging in concentration from 2”5 to 640 pg/ml. Ampicillin was present to select for plasmid retention. Bacteria

grown

(e) Preparation

of

rRNA

and

reverse

transcription

Cells in exponential phase growth were lysed by 3 cycles of freeze-thaw in 150 fig lysozyme/ml in 25 mM-Tris HCl (pH 7.6), 60 m&r-KCl, 10 miv-MgCl,, and 20% (w/v) sucrose. Sodium deoxycholate and Brij-58 were added to 0.2% (w/v) and 0.6% (v/v), respectively, as was 10 pg of DNase (Lindahl & Porchhammer, 1969). Lysates were

Mutagenesis

Near the Decodim

Table 1 Lethaiity

Plasmid pKK3535* pKK 1409U* pNOl409A pNOl409G pNOl491A pN009A-91C: pBOOSA-9lA

of certain, rRNA

FIB101

HBlOl

41 41 5 63 51 62

mutations

+pC1857

Transformation (%) 100 80 3 1 3 5 6

54 1342 385 2209 1082 1062

The plasmids were prepared from E. coli also containing pC1857 (except where indicated *) and used to transform both HBlOl and IIBlOl containing the cZ857 repressor (+pCI857). Numbers of transformants recovered have been adjusted for differences in transformation frequency between the 2 strains (using the transformation of pKK3535). Percentage transformation is the percentage of transformants recovered in HBlOl versus that recovered in HBlOl + pCI857. All HBlOl transformants tested among the latter 5 plasmids were found to contain both pN02680 and pC1857. With the exception of pKKl409U, all results shown are a compilation of at least 3 experiments. cleared by centrifugation at 10,000 revs/min for 15 min. Total cellular RNA was prepared from lysates by extraction with phenol and precipitation with ethanol. Alternatively, lysates were fractionated on a sucrose gradient and REA isolated from subunits and 70 S ribosomes. Nucleic acid (2 mg) was loaded on a 15 ml, 10% to 30% (W/V) sucrose gradient containing 6 mivr-MgCf,, 60 mM-KCl, 20 mm-Tris HCI (pH 7.6): 1 mM-dithiothreitol. Gradients were spun in an SW28 Beckman rotor for 16 h at 17,000 revs/min at 3°C. Fractions of 30 S and 70 S particles or polysomes were pooled, the RNA was precipitated with ethanol, extracted with phenol and reprecipitated. Approx. 5 pmol of 16 S rRNA was used for each primer extension reaction. Primer extension methodology has been described (Tapprich et al., 1989). Quantification of

mutant RNA was done by scanning mutant and wild-type sequencing lanes using an LKR densitometer. Where false banding occurred, an average peak area from the false banding was subtracted from an average peak for a true band and from the mutant or wild-type band. This adjust,ed mu&ant Peak area was expressed as a percentage of an adjusted true band. The percentage increase in a mutant hand and decrease in a wild-type band was determined for each set of reactions.

3. Results (a) Construction

of mutant

rDNA

Mutations were produced at positions 1409 and 1491 in E. co& 16 S rRNA by site-directed mutagenesis (Zoller & Smith, 1984) using the modifications of Kunkel (1985). Eleven of the 15 possible mutations were constructed in M13mp19 and cloned into expression vectors pKK3535 or pN02680, both of which contain the rrnB operon. Expression of rDNA from pKK3535 is controlled by the natural rrnB promoters PI and P2, while vector pN02680 contains the repressible promoter I?,, which can be controlled by a temperature-sensitive repressor, ~1857. The EcoRIGXbaI fragment containing 16 S rDNA was cloned back to Ml3 for sequencing to

Region

qf 16 S rRNA

129

assure that the mutations had indeed been cloned. In producing the first three single mutants at 1409, the entire 770 base-pair ApaJ-XbaI fragment, cloned from Ml3 to expression vectors, was sequenced to ensure that no second site mut,ation had been introduced during the mutagenesis. Second mutations were not found. Thus, onl,y the 100 bases surrounding the mutations were analyzed in later mutant production. The 11 mutants produced here were not specifically chosen, but were rather those that appeared first during the mutagenesis. Unexpected difficulties were encountered in unsuccessful attempts to isolate the last four mutants. (b) Lethality

of some rRNA

~~,~tat~~~s

into of certain rDNA mutations Cloning pKK3535 proved to be impossible, indicating that the mutations were lethal in E. co&. The mutations could be propagated in pN02680 in the presence of the repressor on pCI857 at the permissive temperature (30°C). To demonstrate conclusively the lethality of these mutations, plasmids were prepared under repressed conditions and used to transfform E. coli in either the presence or absence of ~1857. Transformations were done at 30°C and the transformation frequencies of mutant and wild- type plasmids were compared in the two strains. A non-lethal plasmid was expected to transform the tvvo strains with approximately equal efliciency. Results shown in Table 1 demonstrate that five of the 11 mutations could be propagated only when in the presence of pC1857. The starting material for these transformations necessarily contained both plasmids, because plasmids containing putative lethal mutations could be maintained only in strains containing pC1857. We were able to slmw that transformants of the five lethal plasmids that did grow in HBlOl without pCI857 had received both plasmids during the transformation. The presence of pC1857 was confirmed by screening for high-level resistance to neomycin ( >50 pg/ml). The percentage transformation of HBlOl of the latter five plasmids in Table 1 represents only the frequency with which a competent cell received both plasmids. We conclude, therefore, that 1409A, 1409G. 1491A, 1409A-1491C and 1409A-1491A are lethal mutations when expressed in E. coli strain HBlOl. (c) Cellular

phenotype

with non-lethal

,mutations

The effect of the six non-lethal mutations on bacterial growth is shown in Table 2. Growth-rates were measured in rich medium (LB) in at least four separate experiments. Doubling times of the different mutants ranged from 64 to greater than 200 minutes, compared to 55 minutes for HBIIOl containing wild-type pKK3535. Transformants with very slow growth-rates reverted to faster growth at a high frequency, accounting for t,he large standard deviations in the doubling times. Care was taken to minimize the number of revertarnt cells by

E. A. De Xtasio and A. E. Dahlberg

130

Table 2 1409-1491 secondary structure versus Piasmid

phenotype

Secondary structure

Growth?

PAR MIC (pg/ml)‘j

C-G G-C c :u Ii-G U-A c:c u:c G:G A:G C:A A:<: A:A

545&l 64f2 71+1 82_+5 85+3 185+33 > 200 lethal lethal lethal lethal lethal

10 10 320 IO 10 320 320

pKK3535 pKKOSG-91C pKK1491U pKK1409U pKKOSU-91A pKK1491C pKKOSU-91C pN01409G pN0 1409A pN01491A pN009A-91C pN009A-91A

Other resistance§ None Low neamine only 9 other aminoglycs NOW

None 9 other aminoglycs 9 other aminoglycs

Compilation of data presented in order of increasing detrimental effect. j’ Doubling time in minutes & 1 standard deviation. 1 Minimal inhibitory concentrations (MIC) of paromomycin. 9 Resistance to other aminoglyeosides (aminoglycs) is shown in Table 4.

tolerated by bacteria containing wild-type plasmid. Resistance to a variety of aminoglycosides is that disrupted shown in Table 3. Mutations Watson-Crick pairing of 1409 and 1491 (C : U, U : C or C : C juxtapositions) yield similar, significant levels of resistance to all of the aminoglycosides that permitted base-pairing shown. Mutations conferred either very low levels or no resistance to most aminoglycosides. All of the viable mut,ations conferred a low level of resistance to neamine. Neamine is composed of only rings I and IT of the four-ring neomycin structure and was the weakest of the neomycins tested. Remarkably, even a simple inversion of the wild-type sequence (C-C to C-C) conferred resistance. Resistances to gentamicin and hygromycin are interesting. In other systems, modiresistance are four bases fications conferring removed from 1409 and 1491, due to methylation at, 1405, and mutation of 1495, respectively (Beauclerk & Cundliffe, 1987; Spangler & Blackburn, 1955). Streptomycin sensitivity was tested by transforming the streptomycin-sensitive strain DHI with

starting all experiments with cells grown on solid agar plates from which revertant colonies had been removed by coring. Coring was achieved by removing from the plate an agar plug containing the entire revertant colony. (d) Aminoglycoside

resistance in vivo

Several mutations at 1409 and 1491 conferred resistance to a variety of aminoglycoside antibiotics. Resistance levels were quantified as the minimal inhibitory concentration (MIC), the lowest drug concentration that inhibited visible bacterial growth. For example, bacteria containing wild-type pKK3535 were inhibited by 10 pg paromomycin/ml but would grow in 5 pg/ml (see Table 2). Plasmids containing 1491U, 1491C or 1409U-1491C, however, conferred resistance such that MIC was raised to 320 pg/ml. The resistant bacteria did not grow indefinitely in these high concentrations of drug; they could, however, be propagated at a drug concentration of two- or fourfold higher than that

Table 3 Aminoglycoside resistance Plasmid pKK3535 pKK3535 pKKOSG-91C pKKOSU-91A pKK1409U pKK1491U pKKOI)U-SIC pKK1491C

Secondary structure C-G C-G G-C U-A U-G c:u u:c C:C

Neo

Par

Rib

Nea

Kan

Tobra

Gent

5 1 1 2 1

10 1 1 1 1

IQ 1 1 1 2

40 1 4 2 2

10 1 2 2 2

25 I 1 1 1

2.5 P 2 2 1

32 32 32

32 32 16

32 32 32

4

16

4

16

16 16

4

16

8

Apra

ffYP 80 1 ND ND 2

I6

10 1 2 2 2 64

16 8

64 32

ND 4

4

Resistance to aminoglycosides conferred by the non-lethal mutations in 16 S rRNA. MIC values of each drug for cells containing pKK3535 are given in line 1; these are assigned a value of 1 (line 2, pKK3535) and the relative increase in MIC for bacteria containing mutant plasmids is shown. Antibiotics are: Neo, neomycin; Par, paromomycin; Rib, ribostamycin; Nea, nearnine; Kan, kanamycin; Tobra, tohramycin; Gent, gentamicin; Apra, apramycin; Hygro, hygromycin. ND, not determined; the drug was not available.

131

Mutagenesis Near the Decodisg Region of 16 S rRNA (f) Analysis

Table 4 rRNA expression

Levels of mutant Plasmid

Lysate

30 s

70 s

Polysomes

pKKl409U pNQ1409G pNOlPO9A pKKO9GSlC pKKi409U-R

55+19 69 36 ND 13

55&S 51* 1 4254 ND ND

54+11 1857 1423 66 ND

53f15 ND ND ND ND

The amount of mutant rRNA is given as a percentage of deviation is indicated where made. pKK1409U-R refers expressed in a revertant strain

present in the indicated fractions the total rRNA. One standard more than 1 determination was to plasmid pKK1409U being of HBlOl. ND, not determined.

mutant plasmids. None of the mutants showed detectable resistance to streptomycin or kasugamycin in this strain. Expression of mutants in DHl and HBlOl gave essentially the same resistance profile for the other aminoglycosides (data not shown).

(e) Utilization

of mutant

ribosomal RNA

Examples of lethal and non-lethal mutants, the 1409, single mutan&, were selected to investigate the degree to which mutant rRNAs were being utilized by the cell. To determine the amount of mutant RNA in ribosomes, lysates were subjected to sucrose density-gradient centrifugation and the RNAs from pooled fractions of 30 S and 70 S peaks were sequenced by reverse transcriptase. The wildtype and mutant lanes were scanned densitometritally to quantify both the reduction in wild-type and the increase in mutant nucleotides as compared to several neighboring bands on the autoradiogram. The results are shown in Table 4. Ribosomal RNA containing the non-lethal mutant 1409U was present in approximately equal amounts in cellular RNA, 30 S, 70 S and polysome fractions, indicating that ribosomes containing this mutant rRNA function in translation. Chemical footprinting of 70 S ribosomes containing 1491U or B491C published previously (De Stasio et al., 1989) showed that these two mutations could be found in significant amounts in 70 S ribosomes. Note, however, that the presence of any of these three mutations retards the bacterial growth-rate (Table 2). Lethal mutants 1409G and 1409A were present in similar amounts in total cellular RNA and the 30 S fractions, but comprised only 14 to 18% of the 16 S rRNA in the 70 S ribosomal fraction. One might expect there to be a paucity of 70 S ribosomes or excess 50 S subunits in these ribosome preparations due to the large percentage of mutant rRNA being selectively omitted from the 70 S population. This was found not to be so, as determined by quantitative scanning of sucrose gradient profiles. No differences in the relative amount of free subunits verSu$ 70 S ribosomes were noted for strains containing wild-type or mutant plasmids (data not shown).

of revertants

Mutant-containing bacteria with very slow growth-rates (e.g. 1491C or 1409U) reverted to faster growth with a frequency of approximately 8 x lop4 per generation. Reversion was most evident on agar plates where larger, faster-growing colonies could be seen growing on a lawn of slowgrowing mutant colonies. Revertants couid be eiiiminated from paromomycin-resistant strains (such as those containing 1491C) by growing the cells on a plate containing paromomycin (10 pg/ml), since revertants not resistant to were this aminoglycoside. A careful analysis of revert,ant strains showed that reversion was due to a host cell mutation. Plasmids prepared from purified reverhant ‘cells showed no evidence of deletions or rea,rrangem’ents in restriction digests. Sequencing confirmed that’ the plasmids still contained the original mutation. In addition, when such plasmids were used to re-transform HBlOl, the majority of transformants had the original slow-growth phenotype, which would again give rise to revertants. Revertant bacteria were analyzed after curing of the plasmids in the absence of selection. The cured revertants were transformed with the original mutant plasmid as well as a variety of other plasmids that conferred a slowgrowth phenotype in HBlOl. The transformants exhibited the faster growth-rate characteristic of revertant cells and similar to wild-type HBIOI containing wild-type plasmid (see Table 5). We therefore conclude that one or more chromosomal alterations are responsible for the majority of these reversions, and that they affect expression of plasmid-coded rDNA by some undetermined mechanism. A reduction in plasmid copy number in the revertant strains has been investigated. The level of ampicillin resistance conferred by pKK3535 in revertant cells was eightfold lower (from 24 to 3 mg/ml in original isolates) than that of pKK3535 in wild-type HBIOl. Similar decreases have been reported by others for reductions in plasmid copy number at lower levels of resistance (e.g. Uhlin & Nordstriim, 1977). The level of mutant rRNA expression, as measured by reverse transcription of total cellular rRNA, revealed a drop from 55% (in wild-type HBlOl + pKK1409U) to 13 y’ mutant 1409U in the revertant strain. The decrease in

Table 5 Growth-rates Strain HBIOI-revertant HBlOl-revertant HBlOl-wild type HBlOl-wild type

of revertant Plasmid

straiss Doubling

pKK149lC pKK1409U pKK1491C pKK1409U

Doubling times +l standard deviation plasmids in wild-type and revertant bacteria.

time (min)

56+7 57k4 185k33 8225 -are sbownl for

2

E. A. De Xtasio and A. E. Dahlberg expression of mutant rRNA can account for the reversion of a wild-type phenotype in both growth rate and paromomycin sensitivity.

4. Discussion (a) Xtructure-phenotype

relationships

Twelve of 16 possible nucleotide juxtapositions at position 1409.1491 in 16 S rRNA have been studied. The variety of phenotypes expressed by these mutations in vivo indicates that changes of rRNA structure in this region have profound effects on ribosome function. Correlations between mutant rRNA structure and the observed bacterial phenotypes are seen in Table 2. It is clear from these data that the mutations at 1409-1491 that are most detrimental to E. eoli contain either two purine bases (G : G, A : G, A : A) or a purine and pyrimidine base that cannot pair (A : C, C : A). These mutations are lethal in plasmid pKK3535, in contrast to mutations in which nucleotides at 1409-1491 can base-pair (C.G, G. C, G .U and U. A) or are both pyrimidine bases (6. C, U. C, C. U). The altered dimension of the 1409-1491 region by bulky or unpaired purine bases would appear to be more critical to the proper tertiary structure (and function) of this region than is base-pairing. It is difficult to find a correlation between structure and growth-rate among the non-lethal mutations. Table 2 shows that a simple disruption of base-pairing is not always correlated with slow growth. Mutations containing C: C or U : C have a dramatic effect in retarding growth (doubling times of 185 and >200 min) in contrast to C : U, which confers a doubling time of 71 minutes. Clearly, the C : U mutation is not as detrimental to the cell as are the other two mutations, although they all contain unpaired pyrimidine bases at 1409-1491. The loss of base-pairing at 1409-1491 confers resistance to a variety of aminoglycoside antibiotics. Mutations that confer resistance include those producing C : U, U : C and C : C juxtapositions at 1409-1491. Using chemical footprinting, we showed (De Stasio et al., 1989) that resistance to paromomycin is strictly correlated with a disruption of the antibiotic-rRNA interaction. Interestinglyi a low level of resistance to neamine and kanamycin is seen with mutations that do contain a base-pair (G. C or U. A). This indicates that the tertiary structure of the region, the presumed antibiotic binding domain, is altered sufficiently even with a C-G to G-C inversion, such that the antibiotics no longer interact as well with the mutant tRNA as they do with wild-type rRNA in the ribosomes. It is important to note that resistance does not correlate simply with the slow-growth phenotype (e.g. compare U-G and C : U mutants, Table 4). Nucleotides in reverse juxtapositions at 1409 and 1491 do not produce equivalent phenotypes. As mentioned, C : U and U : C confer growth rates that differ by over 100 minutes. Even paired nucleotides

confer a different phenotype when their order is reversed. For example, the mutant G-C confers a doubling time that is ten minutes slower than that of the wild-type C-G pair. These data, together with the aminoglycoside-resistance results, pointed out the remarkable sensitivity of this region to subtle changes in both primary and secondary structure, changes in structure that in turn affect ribosome function in a variety of ways, yielding the various phenotypes observed in this study. Structural changes in the rRNA in two of these mutations have been documented (De Stasio et al., 1989). Chemical probing data indicated that 70 S ribosomes containing C: U or C : C mutations bad decreased reactivity at 1408 and that the normally unreactive N3 of 1409 became reactive. In the C : C mutant, 1491 and 1494 were altered in their reactivity. Lastly, the more distant bases, 1418 and 1483, were more reactive in ribosomes containing either mutation, indicating relatively long-range changes in RNA structure. In contrast, ribosomes containing 1409U (U-G) had a pattern of reactivity indistinguishable from wild-type (C-G) ribosomes. These data correlate precisely with the aminoglycoside-resistance phenotype. The U-G mutation, like the wild-type, C-G: confers no resistance, while the other two mutations (C : C and C : U) do. The chemical probing data do not, however, correlate with growth-rate data. The U-G mutant grows slower than does C: U and yet the C : U mutation causes greater structural alterations and yields antibiotic resistance. The C : U mutation has a reactivity pattern indistinguishable from C : C-containing ribosomes, although the latter mutation is much more detrimental to bacterial growth. These data indicate either that primary structure is very important or that structurally undetectable but functionally important changes have occurred in the higherorder structure of the mutant rRNAs. (b) Function

of mutant

rRNA

This study demonstrates the diversity of phenotypic expression exhibited by the 12 variants sf position 1409 and 1491 in 16 S rRNA, and underscores the need to construct a variety of changes at a base-paired site when evaluating its functional importance. That none of the mutations is silent indicates that the integrity of this position of 16 S rRNA is very important for proper ribosome structure and function. Phenotypically silent mutations in the 1400 region of 16 S rRNA have been reported (Jemiolo et al., 1985; Thomas et al., 1988; Denman ei: al., 1989), indicating that not, all mutations in the decoding region produce detectable alterations in ribosome function. The mutations at positions 1409 and 1491 have been characterized by viability, growth-rate and antibiotic resistance. The next step is to determine the precise effects of these mutations on ribosome function. That the mutants are being expressed in the cells is obvious from the fact that they confer phenotypes, particularly antibiotic resistance, and it is supported by chemical probing

133

Mutagenesis Near the Decoding Region of 16 X rRNA and sequence analysis of the rRNAs in 30 S subunits, 70 S ribosomes and polysomes. As noted by others (e.g. Vester & Garret, 1988), it is curious that mutant ribosomes that make up approximately SOo/o of the population of cellular RNA can have a dominant lethal or slow growth effect. It is apparent that phenotypic dominance of mutant rRNA is dose-dependent. Paromomycin resistance and slow growth are seen when mutant rRNA represents 56o/o of the total rRNA. Dominame of both phenotypes is lost in revertant cells that contain only 13% mutant rRNA. Finally, we have shown that rRNAs containing lethal mutations are selectively excluded from 70 S ribosomes. Rottman et al. (1988) have made a similar observation with a lethal double mutation at 1399 and 1401. Expression of this mutant rRNA, however, leads to an excess of 50 S subunits. Thus, variable fates await the different lethal mutant rRNAs, although the final outcome is always the same.

Denman, R., Negre, D., Cunningham, I’. R., Nurse, K., Colgan, J., Weitzmann, C. & Ofengand, J. (1989). Biochemistry, 28, 101221019. Gourse, R. L., Takebe, Y., Sharrock, R. A. & Nomura, M. (1985). Proc. Nat. Acad. Sci., U.S.A. 82. 1069-1073. Gutell, R. R., Weiser, B.; Woese, C. R. & Nosller, H. F. (1985). Progr. Nucl. Acids Res. MOE. Biol. 32, 155-216. Hanahan, D. (1983). J. Mol. Biol. 166, 557-580. Hui, A. & de Boer, H. A. (1987). Proc. Nat. Acad. Sk., 1J.S.A. 84, 4762-4766. Hui, A., Eaton, 2). H. & de Boer, H. A. (1988). EMKO J. 7, 438334388. Jacob, W. F., Santer, M. & Dahlberg, A. E. (1987). Proc. Nat. Acad. Sci., U.S.A. 84, 4757-4761. Jemiolo, D. K., Zwieb, C. & Dahlberg, A. E. (1985). Nucl.

We thank B. T. De Stasio, H. U. Goringer, D. Jemiolo,

Murgola, E. J., Hijazi, K. A., Goringer. H. lJ. & Dahlberg, A. E. (1988). Proc. Nat. Acad. Sci., U.S.A. 85, 4162-4165. Remaut, E., Stanssens, P. & Fiers, W. (19811. i&x, 15, 81-93. Rottman. N., Kleuvers, B., Atmadja, J. & Wagner, R.

Acids Res.

13, 8631-8643.

Kunkel, T. (1985). Proc. Nut. Acad. Sai., U.S.A. 82, 488-492. Lindahl, L. & Forchhammer, J. (1969). J. Mol. Riol. 43, 593-606. Moazed, D. & Noller, H. F. (1986). Cell, 47, 9855994. Moazed, D. & Noller, H. F. (1987). Nature (London), 327, 389-394.

P. Makosky. C. Prescott, W. Tapprich, G. Q. Pennable and Melvin Santer for helpful discussions, and A. Landy for the use of the densitometer. This work was supported by NIH grant GM19756 to A.E.D. and an NIH training grant T32GM07601 to E.A.D.

(1988). Eur.

J. Biochem.

177, 81-90.

Sanger, F., Nicklen, S. & Coulsen, A. R. (1977). Proc. Nat. Acad. Sci., U.S.A.

Beauclerk,

A. A. D. & Cundliffe, E. (1987). J. Mol. Biol. 193, 661-671. Bolivar, F. & Baehman, H. (1979). Methods Enzymol. 68, 2455267. Brosius, J., Ullrich, A., Raker, M. A., Gray, A., Dull, T. J.; Gutell, R. R. & Noller, H. F. (1981). Plarmid, 6, 112-1;18.

Dahlberg, A. E. (1989). CeEZ,57, 525529. De Stasio, E. A., Goringer, H. U., Tapprich, W. E. & Dahlberg, A. E. (1988). In Genetics of Translation (Tuite, M. F., Pieard, M. & Bolotin-Fukuhara, M., eds), pp. 17-41, Springer-Verlag, New York. De Stasio, E. A., Moazed, D.. Noller, H. F. & Dahlberg, A. E. (1989). EMBO J. 8. 1213-1216.

74, 5463-5467.

@angler, E. A. & Blackburn, E. A. (1985). J. ,BioZ. Chem. 260, 6334-6340. Stern, S., Weiser, B. & Noller, H. F. (1988). J. Mol. Biol. 204, 447-48 I. Tapprich, W. E., Goss, D. J. & Dahlberg, A. E. (1989). Proc. Nat.

Acad

Sci.,

U.S.A.

86, 4927-4931.

Thomas, C. L., Gregory, R. J., Winslow, G., Muto, A. & Zimmermann, R. A. (1988). NucZ. Acids Res. 16, 8129-8146. Uhlin, B. E. & Nordstrom, K. (1977). Plasmid, 1, l--7. Vester, B. & Garrett, R. A. (1988). EMBO J. 7, 357773587. Zoller, M. J. & Smith, M. (1984). DNA, 3, 4X-488.

Edited by N. Sternberg