A Novel Insertion Mutation in Streptomyces coelicolor Ribosomal S12 Protein Results in Paromomycin Resistance and Antibiotic Overproduction

Abstract

We identified a novel paromomycin resistance-associated mutation in rpsL, caused by the insertion of a glycine residue at position 92, in Streptomyces coelicolor ribosomal protein S12. This insertion mutation (GI92) resulted in a 20-fold increase in the paromomycin resistance level. In combination with another S12 mutation, K88E, the GI92 mutation markedly enhanced the production of the blue-colored polyketide antibiotic actinorhodin and the red-colored antibiotic undecylprodigiosin. The gene replacement experiments demonstrated that the K88E-GI92 double mutation in the rpsL gene was responsible for the marked enhancement of antibiotic production observed. Ribosomes with the K88E-GI92 double mutation were characterized by error restrictiveness (i.e., hyperaccuracy). Using a cell-free translation system, we found that mutant ribosomes harboring the K88E-GI92 double mutation but not ribosomes harboring the GI92 mutation alone displayed sixfold greater translation activity relative to that of the wild-type ribosomes at late growth phase. This resulted in the overproduction of actinorhodin, caused by the transcriptional activation of the pathway-specific regulatory gene actII-orf4, possibly due to the increased translation of transcripts encoding activators of actII-orf4. The mutant with the K88E-GI92 double mutation accumulated a high level of ribosome recycling factor at late stationary phase, underlying the high level of protein synthesis activity observed.

The 30S ribosomal subunit in bacteria is responsible for decoding the genetic information carried by mRNA and, in cooperation with the 50S subunit, participates in peptidyl translocation (5). Recent progress in ribosomal structure research has unraveled the molecular mechanisms underlying the inhibitory actions of antibiotics on the protein biosynthetic machinery (2, 6, 19, 36, 39, 40). For example, streptomycin and paromomycin, which belong to the streptamine- and 2-deoxystreptamine-containing groups, respectively, are aminoglycoside antibiotics that cause miscoding and that have distinct chemical structures (Fig. 1). Although both bind to the decoding center, they bind to the 30S subunit at different sites, which causes the ribosome to select incorrect aminoacyl-tRNAs (aa-tRNAs) and which leads to misreading during translation.

FIG. 1.

Chemical structures of paromomycin and streptomycin.

Paromomycin, an antibiotic used to treat intestinal infections such as cryptosporidiosis and amoebiasis, also interferes with protein translation by altering the initial tRNA selection. The binding of paromomycin to helix 44 (h44) in the A site of 16S rRNA induces adenosines 1492 and 1493 (in the Escherichia coli numbering) to flip out of the h44 groove, similar to the conformational change observed when a cognate aa-tRNA enters the A site (2, 25). Moreover, paromomycin stabilizes these conformational changes, reducing the energetic cost threshold for a nearly cognate aa-tRNA to enter the A site (2, 25). Paromomycin has also been reported to inhibit translocation by increasing the affinity of aa-tRNA for the A site (29), to stimulate 70S ribosome association (8), and to prevent peptide release (42). Most paromomycin resistance mutations reported to date are located in 16S rRNA (3, 9, 18, 30, 31, 34).

Ribosomal protein S12, which is located at the interface of two subunits near the decoding center of the ribosome, is important in maintaining translational accuracy (2, 32, 36, 39). Several S12 amino acid residues help with the positioning of 16S rRNA nucleotides during the codon-anticodon interaction. Aside from drug resistance, many (but not all) S12 mutations show other pleiotropic effects, including hyperaccuracy, a reduced growth rate, poor support for bacteriophage growth, and impaired peptide chain elongation (22, 43). Working with Streptomyces coelicolor A3(2), we have shown that the K88E (K87 in E. coli) amino acid substitution in the S12 protein confers a high level of resistance to streptomycin and activates the production of the blue-colored polyketide antibiotic actinorhodin (Act) (7, 33). Later, we found that ribosomes containing the K88E mutant S12 protein sustain a high level of protein synthesis activity during late growth phase, resulting from the increased stability of the 70S ribosomal particle and the enrichment of some translation factors, such as ribosome recycling factor (RRF) (11, 27).

We have also described a novel paromomycin resistance mutation in ribosomal protein S12, consisting of amino acid substitution P91S (P90 in E. coli), which increases the level of paromomycin resistance eightfold and which leads to Act overproduction (26). During our recent studies of cumulative drug resistance mutations in S. coelicolor, we identified another paromomycin resistance mutation in the S12 protein (38). Our ultimate aim is to develop ribosome engineering as a rational approach to taking full advantage of the capabilities of bacterial systems (23, 24). Here, we show that this novel insertion mutation causes an increase in the level of paromomycin resistance of over 20-fold and, in combination with the K88E mutation, the significant overproduction of Act and another red-colored antibiotic undecylprodigiosin (Red).

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The prototropic wild-type S. coelicolor A3(2) strain 1147 and its derivatives used in this study are listed in Table 1. GYM, R5MS, SFM, and R2YE agar media were prepared as described previously (16, 21). YEME liquid medium (16) was used to grow cells for ribosome preparation and to assay for Act production in liquid culture. Spontaneous paromomycin-resistant mutants were obtained by spreading spores of wild-type strain 1147 or K88E mutant strain KO-178 (7) onto GYM plates containing various concentrations of paromomycin (3 to 10 μg/ml). All Streptomyces strains were cultured at 30°C.

TABLE 1.

S. coelicolor A3(2) strains used in this study

Strain	Descriptiona	Mutation in rpsL geneb	Amino acid substitution(s) or insertion(s)	Frequency of mutationc	Reference or source
1147	Prototrophic wild type	—d			16
KO-178	Smr mutant from 1147 (str-18)	A262 → G	K88E	NDe	7
KO-347	Parr mutant from 1147 (par-1)	C271 → T	P91S	ND	26
92G	Parr mutant from 1147 (par-2)	274-276::GGG	GI92f	3/25	This study
SP1	Parr mutant from KO-178 (str-18, par-1)	A262 → G, C271 → T	K88E, P91S	10/30	This study
SP2	Parr mutant from KO-178 (str-18, par-2)	A262 → G, 274-276::GGG	K88E, GI92	3/30	This study
KO-948	GI92 gene replacement strain of 1147	274-276::GGG	GI92	ND	This study
KO-949	K88E GI92 gene replacement strain of KO-178	A262 → G, 274-276::GGG	K88E, GI92	ND	This sutdy

^aSm, streptomycin; Par, paromomycin.

^bNumbering from the first nucleotide of start codon (GTG) of the open reading frame.

^cNumber of strains containing the corresponding mutation/total number of paromomycin-resistant isolates sequenced.

^d—. wild type, no mutation.

^eND, not determined.

^fGlycine insertion at position 92.

DNA sequencing.

The S12-encoding rpsL gene was amplified by PCR from the E. coli genome by using the primers described previously (10) and from the S. coelicolor genome with primers rpsL-F (5′-ATTCGGCACAGAAACCGGAGAAG-3′) and rpsL-R (5′-AGAGGAGAACCGTAGACCGGGTCGA-3′). The amplification protocol consisted of an initial denaturation at 96°C for 3 min, followed by 30 cycles of denaturation at 98°C for 10 s and amplification at 60°C for 1 min, with a final extension at 72°C for 5 min. The purified PCR products were sequenced with a BigDye Terminator Cycle sequencing kit (Perkin-Elmer Applied Biosystems, Foster City, CA).

RT-PCR and real-time qPCR analysis.

Total RNA was prepared from cells grown in YEME liquid medium as described previously (21). The transcription of actII-orf4 was analyzed by reverse transcriptase PCR (RT-PCR) with primers orf4-F (5′-ACCGATGCGGGATGTGTAATTCCG-3′) and orf4-R (5′-GTGCGCGATATTGCTTTCGAGCAC-3′) and a Thermscript RT-PCR kit (Invitrogen), according to the manufacturer's instructions. One microgram of each total RNA was used for the reverse transcription reaction after it was treated with RNase-free DNase I (Invitrogen). PCR amplification (50 μl) was for 26 cycles with 1 μl of the reverse transcription product as the template, and 8 μl of each PCR product was electrophoresed on 1% agarose gels.

Real-time quantitative PCR (qPCR) analysis of gene transcription was conducted with a model 7300 real-time PCR system and Power SYBR green PCR master mixture (Applied Biosystems), as described previously (38). The transcription of the 16S rRNA gene was used as an internal control. The level of transcription in each assay was normalized to the corresponding level of transcription of 16S rRNA. The following primers were used: primers orf4-F2 (5′-TGATCGACGAGGACGAACTCG-3′) and orf4-R1 (5′-ATTCGCGTCGATACGGACCTG-3′) for actII-orf4 and primers 16S-F1 (5′-GCGATAGCCTGATGCAGCGACG-3′) and 16S-R (5′-GCGCATTTCACCGCTACACCAGG-3′) for the 16S rRNA gene.

Gene replacement of rpsL.

Allele exchange was carried out as described previously (21, 41), except for the use of pKC1139, which carries a pSG5-derived temperature-sensitive replicon and which cannot replicate at 37°C (16). A ca. 1.3-kb region that contained approximately 300 bp upstream and 600 bp downstream of the rpsL gene was amplified from the total DNA of S. coelicolor strains 92G and SP2 with primers S12-600-BamF (5′-TTGGATCCCCTACTTCGTCCGCCACG-3′) and S12-600-BamR (5′-TTGGATCCGTCGTCTTGCCCGCGTCG-3′) (the BamHI sites are underlined). The PCR products were cloned into the HincII site of pUC18. After sequence confirmation, the rpsL gene fragments were subcloned into the BamHI site of pKC1139 to yield pKC1139-92G and pKC1139-SP2, respectively. Recombinant plasmids pKC1139-92G and pKC1139-SP2 were introduced by conjugation into S. coelicolor strains 1147 and KO-178, respectively, and transformants were selected with apramycin. The two transformants of each of strains 1147 and KO-178 were cultured on apramycin-containing plates at 37°C, followed by two rounds of nonselective culture at 37°C to allow the loss of the vector. Thereafter, serial dilutions of the resulting spores were plated and incubated at 30°C, and colonies sensitive to apramycin were selected. The correct replacement of the rpsL gene with the mutant allele was confirmed by PCR and DNA sequencing.

Determination of MICs and antibiotic production.

MICs were determined by spotting spore solutions (∼10⁶) onto drug-containing GYM plates, followed by incubation for 48 h at 30°C. The level of Act production on the plates was assessed directly by determination of the intensity of the blue color. For Act production in liquid medium, 0.5 ml of each culture was treated with 1 M KOH and centrifuged at 3,000 × g for 5 min; the optical density at 640 nm (OD₆₄₀) of the supernatant was then measured (16). Likewise, the level of Red production was determined as described by Kieser et al. (16).

Preparation of ribosomes and S-150 fraction.

Cells grown in YEME liquid medium were harvested by filtration at midexponential and stationary phases (phases S1 and S2) (Fig. 2D) and were washed twice with standard buffer (10 mM Tris-HCl [pH 7.7], 10 mM magnesium acetate, 30 mM ammonium acetate, 1 mM dithiothreitol) containing 1 mM phenylmethylsulfonyl fluoride (PMSF). The cell pellets were stored at −80°C until use. The pellets were grinded with aluminum oxide powder (3 g per 1 g cells; Wako) for 15 min at 4°C and suspended in standard buffer containing 1 mM PMSF plus 10% glycerol and 10 U/ml of RNase-free DNase I (Takara). After incubation on ice for 10 min, the solution was centrifuged at 30,000 × g for 30 min at 4°C to remove the cell debris. The supernatant was then fractionated into the ribosome fraction (precipitate) and the S-150 fraction (supernatant) by centrifugation at 150,000 × g for 3 h. The ribosomes were washed once with standard buffer plus 10% glycerol. Following extensive dialysis (twice for 3 h each time) against 60 volumes of standard buffer plus 10% glycerol, the ribosomes and S-150 fraction were stored, in small aliquots, at −80°C until use.

FIG. 2.

Growth, levels of Act and Red production, and levels of actII-orf4 transcription in wild-type and mutant strains. (A) Act production by various mutant strains on R5MS agar plates after 5 days. The tops of the plates are shown to illustrate the production of the blue-colored antibiotic Act and the morphology of each strain. (B) Act production in YEME liquid medium. Fresh spore suspensions (ca. 10⁹ spores) were inoculated into 100 ml YEME liquid medium in a 500-ml flask and incubated on a rotary shaker (200 rpm). Three flasks were used for each strain, and the level of Act production is expressed as the mean value of Act production (OD₆₄₀)/unit cells (OD₄₅₀). WT, wild type. (C) Red production on GYM agar medium. Strains were grown on GYM agar medium covered with a cellophane sheet for 2 or 4 days. The levels of Red produced by the cells were assayed as described previously (16) and are expressed as nmol/mg cell dry weight. (D) Growth in liquid medium. Cultivation was performed as described above for panel B. The A₄₅₀ was monitored, with an A₄₅₀ of 0.06 defined as the zero time point after incubation for 16 to 24 h (to synchronize the display of the growth curves shown here). Growth phases are indicated as follows: E, midexponential phase; T, transition phase; S1, early stationary phase; and S2, late stationary phase. (E) Transcriptional analysis of actII-orf4 by real time-qPCR. Total RNA preparation and real time-qPCR were performed as described in Materials and Methods. The transcription of each gene was normalized to that of 16S rRNA. The error bars indicate the standard deviations of the means of triplicate samples. T, S1, and S2 are growth phases and are as defined above for panel D.

In vitro protein synthesis.

The cell-free synthesis of green fluorescent protein (GFP) was performed as described previously (11). Briefly, a 100-μl reaction mixture containing 20 A₂₆₀ units/ml of ribosome and 50 mg/ml of the S-150 fraction were preincubated at 30°C for 15 min, and GFP synthesis was initiated by adding 100 μg of gfp mRNA. Aliquots (10 μl) were withdrawn at the indicated times, separated on native 10% polyacrylamide gels, and subjected to fluorescence intensity analysis with a FluoroImager (Molecular Dynamics). In some cases, paromomycin was included before the preincubation.

In vitro misreading assay.

To measure the amino acid misincorporation rates, we took advantage of the fact that the GFP protein has only one tryptophan residue (Trp58) in its primary sequence. The deacylation of aa-tRNA was performed as described previously (28). In vitro GFP synthesis was carried out essentially as described as above, except that (i) Trp was omitted from the amino acid mixture and (ii) aliquots were taken every 60 min for 3 h. Determination of the level of GFP synthesis in the presence of Trp (used as a control) was also carried out at the same time. The GFP synthesis rates were linear, at least within the first 2 h, irrespective of the presence or absence of Trp. The in vitro misreading level was expressed as the GFP synthesis rate in the absence of Trp/the GFP synthesis rate in the presence of Trp. GFP synthesis was confirmed also by Western blotting analysis, the results of which were consistent with those of the fluoroimage analysis.

Western blotting analysis.

Blotting of protein onto polyvinylidene difluoride membranes (Immobilon-P; Millipore) and development of the blots with an ECL Western detection system (Amersham) were performed as described previously (11). For RRF analysis, S-150 fractions (8 μg) prepared from cells harvested at stationary phases S1 and S2 were used. Polyclonal anti-RRF antibody was used at a dilution of 1:10,000, while anti-GFP antibody (Roche) was diluted 1:3,000.

RESULTS

Isolation of a novel paromomycin resistance mutation.

A certain paromomycin resistance mutation (P91S in S12) had been shown to activate Act production (26). Thus, during our studies of the effects of cumulative drug resistance mutations on the activation of Act production in S. coelicolor, we introduced paromomycin resistance mutations into a mutant with triple mutations, mutant SGR (14). Mutant SGR contains a K88E mutation in S12, which confers resistance to streptomycin, as well as mutations conferring resistance to gentamicin and rifampin (rifampicin). Paromomycin-resistant mutants were developed spontaneously on paromomycin-containing plates. Mutants were purified by the isolation of single colonies. To our surprise, the mutants which overproduced especially high levels of Act did not have the P91S mutation; rather, they all harbored a novel S12 mutation, consisting of the insertion of a glycine residue at position 92 due to the insertion of three guanosines in the rpsL gene (38). Hereafter, this insertion mutation is designated the GI92 (glycine insertion at position 92) mutation. On the other hand, the mutants which overproduced a moderate amount of Act were found to have the P91S mutation (38).

To study the novel GI92 mutation in more detail, we introduced this mutation, through similar spontaneous mutant selection, into wild-type strain 1147 and streptomycin-resistant rpsL K88E strain KO-178 to generate mutants with single and double mutations, respectively (Table 1). A mutant containing double mutations, K88E and P91S, was also obtained and was used as a reference. The frequency of appearance of the GI92 insertion mutation was lower than that of the P91S mutation (Table 1).

Physiological characterization of the GI92 insertion mutation.

As expected, the mutant SP2 harboring the double mutations GI92 plus K88E produced abundant Act on R5MS agar plates (Fig. 2A) and in YEME liquid medium (Fig. 2B). Likewise, mutant SP2 produced abundant amounts of the red-colored antibiotic Red on GYM agar plates (Fig. 2C). In contrast, the mutant with the K88E mutation (mutant KO-178) and the mutant with the double mutations K88E and P91S (mutant SP1) produced moderate amounts of Act and Red. Surprisingly, the GI92 insertion mutation itself (i.e., in mutant strain 92G with a single mutation) did not result in the overproduction of Act (Fig. 2A and B), although it was slightly effective at inducing the overproduction of Red (Fig. 2C). RT-PCR analysis indicated that the Act overproduction by mutant SP2 with the double mutation (and by mutant KO-178 with the K88E mutation) was due to the enhanced expression of the pathway-specific regulatory gene actII-orf4 at the transcriptional level (data not shown). This was confirmed by real-time qPCR, in which the level of expression of actII-orf4 in mutant SP2 was two- to threefold higher than that in wild-type strain 1147 (Fig. 2E).

Introduction of the GI92 mutation somewhat reduced the bacterial growth rate; this was especially pronounced when the GI92 mutation was introduced into the K88E genetic background (Fig. 2D). We found that the GI92 mutation conferred a much higher level of resistance to paromomycin than the P91S mutation did, which increased the level of resistance to paromomycin by only eightfold (26) (Table 2). The introduction of a single paromomycin resistance mutation (GI92 or P91S) caused cross-resistance to several aminoglycoside antibiotics, including streptomycin, neomycin, kanamycin, gentamicin, and Geneticin. In the K88E genetic background, however, the introduction of P91S or GI92 resulted in hypersensitivity to kanamycin; neomycin; and certain nonaminoglycoside antibiotics, such as fusidic acid, lincomycin, and tetracycline (Table 2).

TABLE 2.

Levels of resistance of S. coelicolor A3(2) mutants to various drugs

Strain (mutation)	MIC (μg/ml) a
	Sm	Par	Km	Neo	FA	Lin	Tet	Gen	G418
1147 (wild type)	1	0.1	0.8	0.3	50	50	10	0.1	0.3
KO-178 (K88E)	100	0.3	0.8	0.8	50	40	10	0.1	0.3
KO-347 (P91S)	20	0.8	0.5	0.5	50	50	5	0.5	0.3
92G (GI92)	20	2	1	0.5	50	20	10	0.5	1
SP1 (K88E, P91S)	200	2	0.5	0.2	20	40	3	0.5	0.5
SP2 (K88E, GI92)	>200	3	0.5	0.2	20	10	3	0.5	0.5

^aMICs were determined following incubation on GYM plates at 30°C for 48 h. Abbreviations: Sm, streptomycin; Par, paromomycin; Km, kanamycin; Neo, neomycin; FA, fusidic acid; Lin, lincomycin; Tet, tetracycline; Gen, gentamicin; and G418, Geneticin. Boldface indicates a decline in the level of resistance.

To clearly demonstrate that the GI92 mutation is responsible for the phenotype observed (i.e., increased resistance to paromomycin and a greater ability to produce antibiotic in the genetic background with the K88E mutation), we performed a set of gene replacement experiments (see Materials and Methods). The strain with the gene replacement that was constructed, strain KO-949 (which harbored the K88E and GI92 mutations) (Table 1), showed increased resistance to paromomycin, as did mutant strain SP2 harboring the K88E and GI92 mutations (data not shown); and KO-949 showed markedly increased levels of Act production on an R5MS agar plate (Fig. 3). On the other hand, a strain with another gene replacement, strain KO-948 (which harbored the GI92 mutation), showed a phenotype similar to that of mutant strain 92G with the GI92 mutation. These results demonstrate unambiguously the causality of the GI92 mutation for the phenotypes observed.

FIG. 3.

Act production in strains KO-948 and KO-949 with rpsL gene replacements and the corresponding parent strains, wild-type (WT) strain 1147 and KO-178, respectively. The strains were inoculated on an R5MS agar plate and were incubated at 30°C for 4 days. The bottom of the plate was photographed to show the production of the blue-colored antibiotic Act.

SP2 mutant ribosomes resist high concentrations of paromomycin and are error restrictive.

Since paromomycin targets ribosomes, we assayed the resistance of isolated ribosomes to paromomycin. In the presence of 0.2 μg/ml paromomycin, protein synthesis by wild-type ribosomes or ribosomes with the K88E mutation was significantly inhibited in vitro by more than 80%, whereas protein was actively synthesized by mutant ribosomes harboring the GI92 mutation (Fig. 4). The highest level of resistance to paromomycin was shown by SP2 mutant ribosomes, which harbored both the K88E and the GI92 mutations.

FIG. 4.

Resistance of mutant ribosomes to paromomycin. In vitro GFP synthesis in the presence of paromomycin was assayed by using ribosomes prepared from exponentially growing cells with a reaction time of 90 min. Paromomycin was added at a concentration of 0, 0.2, 0.5, 1, or 2 μg/ml. Reaction aliquots (10 μl) were separated on a native 10% polyacryamide gel. The upper panel shows fluorographs of the GFP synthesized. The intensity of the GFP bands was determined by scanning the fluorographs (lower panel). Symbols: ▪, wild-type (WT) strain 1147; ▵, strain KO-178; ○, strain 92G; ▴, strain SP1; •, strain SP2.

We next conducted an in vitro misreading assay using a system based on cell-free GFP synthesis (see Materials and Methods). As shown in Fig. 5, the ribosomes with the K88E-GI92 double mutation, as well as ribosomes with the K88E mutation, were characterized by error restrictiveness, whereas the mutant ribosomes harboring only the GI92 mutation exhibited no changes in the frequency of errors in comparison with that of the wild-type ribosomes.

FIG. 5.

In vitro misreading assay. The assay system and the method used to calculate the misreading level (GFP synthesis rate in the absence of Trp/GFP synthesis rate in the presence of Trp [−Trp/+Trp]) of each strain are described in detail in Materials and Methods. The mean values of two or more separate assays are shown.

SP2 mutant ribosomes exhibit aberrant translational activity.

Recent work in our laboratory revealed that mutant ribosomes harboring S12 proteins containing the mutations K88E (in S. coelicolor), K87E (in E. coli), K88R (in Streptomyces albus), and K56R (in Bacillus subtilis) sustain a high level of protein synthesis activity even during late growth phase (10, 11, 17, 27, 35). We therefore reasoned that the enhanced Act production shown by S. coelicolor SP2 harboring the double mutations K88E and GI92 may be indicative of a similar effect on protein synthesis. To test this idea, we measured the in vitro translational activity of ribosomes isolated from wild-type and mutant cells during late stationary phase (the S2 phase) using a GFP cell-free translation system (see Materials and Methods). Notably, mutant ribosomes isolated from strain SP2 cells during the S2 phase exhibited sixfold greater activity relative to that of ribosomes isolated from the wild-type strain (Fig. 6). In contrast, ribosomes from the mutant with the K88E mutation (mutant KO-178) and the mutant with the double K88E and P91S mutations (mutant SP1) did not show such a dramatic increase in ribosomal activity. Importantly, the presence of the GI92 mutation alone did not result in enhanced protein synthesis activity, indicating that the GI92 mutation can exert its positive effect only in combination with the K88E mutation. Unlike the active protein synthesis that occurred during late growth phase, ribosomes isolated from the SP2 cells with the double mutation displayed a low level of protein synthesis activity during the fast-growing exponential phase (Fig. 4, paromomycin lane 0 μg/ml), accounting for the observed slow growth of strain SP2 (Fig. 2D).

FIG. 6.

In vitro synthesis of GFP by using wild-type and mutant ribosomes prepared from cells in late growth phase. Ribosomes and S-150 fractions prepared from cells grown to stationary phase (S2 phase) (Fig. 2D) were used for determination of the levels of cell-free GFP synthesis. Reaction aliquots (10 μl) were sampled every 40 min. The upper panel shows fluorographs of the GFP synthesized. The intensity of the GFP bands was determined by scanning the fluorographs (lower panel). Symbols: ▪, wild-type (WT) strain 1147; ▵, strain KO-178; ○, strain 92G; ▴, strain SP1; •, strain SP2.

The S-150 fraction from the SP2 mutant cells stimulates protein synthesis in vitro.

To determine whether ribosomes and the S-150 fraction (which contains all the translational cofactors) were both necessary for the observed increase in the level of protein synthesis activity of mutant SP2, we mixed wild-type ribosomes with the SP2 S-150 fraction and we mixed SP2 ribosomes with the wild-type S-150 fraction of cells grown to late stationary phase (S2 phase). The level of GFP synthesis was significantly higher when wild-type ribosomes were incubated with the SP2 S-150 fraction than when they were incubated with the wild-type S-150 fraction (Fig. 7A). Conversely, the level of GFP synthesis was significantly lower when mutant SP2 ribosomes were incubated with the wild-type S-150 fraction than when they were incubated with the SP2 S-150 fraction, indicating a need for the S-150 fraction of mutant SP2 for maximal translation activity. The S-150 fraction from mutant SP2 (and also mutant KO-178) had a greater amount of RRF than the S-150 fraction from the wild type, as determined by Western blotting analysis (Fig. 7B), accounting at least partly for the high level of translation activity observed in mutant strains SP2 and KO-178 (Fig. 6). These results agree with previous findings obtained with the mutant with the K88E mutation (11).

FIG. 7.

(A) Effects on GFP synthesis of cross-mixing of S-150 fractions and ribosomes from wild-type and mutant cells. Ribosomes and S-150 fractions were prepared from wild-type strain 1147 and SP2 cells grown to late stationary phase (S2 phase). The level of cell-free synthesis of GFP was determined as described in the legend to Fig. 6. The upper panel shows fluorographs of the GFP synthesized. The intensity of the GFP bands was determined by scanning the fluorographs and is depicted by the symbols shown in the upper panel (lower panel). (B) Western blot analysis of RRF. The S-150 fraction was prepared from cells grown to the early stationary phase (S1 phase) or late stationary phase (S2 phase) as described above for panel A and subjected to Western blotting. WT, wild type.

No occurrence of GI92 insertion mutation among paromomycin-resistant E. coli isolates.

Certain mutants of E. coli and Thermus thermophilus with mutations in rpsL have been shown to be resistant to paromomycin (1, 20). Therefore, we examined whether or not the GI92 insertion mutation is also found in paromomycin-resistant E. coli isolates. Almost all resistant isolates (63 of 64 isolates tested) of E. coli BW25113 which developed on the plates containing 50 μg/ml paromomycin (10-fold the MIC) were found to have an rpsL mutation, such as P90L, P90Q, or G91V (in the E. coli numbering), but the insertion mutation was not found. These results may be an indication that the relatively frequent appearance of the GI92 insertion mutation (Table 1) is a characteristic of actinomycetes, which are distinguished by a high guanine-cytosine content in their DNA compositions.

DISCUSSION

In this work, we identified a novel paromomycin resistance mutation in ribosomal protein S12 of S. coelicolor caused by the insertion of a glycine residue at position 92 (corresponding to position 91 in the E. coli S12 ribosomal protein). To our knowledge, this is the first report of an insertion mutation in a ribosomal protein that confers resistance to aminoglycosides. This new mutation conferred a higher level of paromomycin resistance than the previously identified P91S mutation in S12 did. Moreover, when this insertion mutation was combined with the K88E mutation, it was responsible for a high level of translation activity during late growth phase. We previously reported that certain ribosomal mutations that confer a high level of translational activity during late growth phase activate the production of antibiotics as well as certain enzymes and induce tolerance to toxic organic compounds (12-15, 17, 33). This new approach to the elicitation of the cell's full capability was called ribosome engineering (23, 24). It is therefore of particular interest that the GI92 insertion mutation alone did not affect the translational activity during late growth phase but did so when it was combined with the K88E mutation. It is also notable that these ribosomal activity patterns (Fig. 6) agreed well with the levels of Act and Red production of each strain (Fig. 2 A to C), confirming our previous proposal (11, 27). The principal regulator of Act production in S. coelicolor appears to be the availability of the pathway-specific transcriptional regulatory protein ActII-ORF4, a threshold concentration of which is required for the efficient transcription of its cognate biosynthetic structural genes (4). Although we do not yet know how the drug resistance mutations mediated preferential gene transcription (Fig. 2E), it is conceivable that the expression of pathway-specific regulatory genes (e.g., actII-orf4 for Act and redD for Red) is governed by higher-order regulatory proteins and that expression of the latter presumptive regulatory proteins may be significantly affected under conditions associated with enhanced protein synthesis during stationary phase in the mutants.

Paromomycin, which contains four rings, including a two-ring (I and II) neamine core (Fig. 1), binds to the groove at one end of h44 of 16S rRNA and interacts with several nucleotides around the A site (2, 37). The amino acids around P90 (P91 in S. coelicolor) of S12 are adjacent to the paromomycin binding site. Thus, insertion of a glycine residue at position 92 might induce local conformational changes that lead to paromomycin resistance by reducing its affinity for the ribosome. Since the GI92 mutation alone did not enhance the cell's protein synthesis activity during late growth phase (Fig. 6), the molecular mechanisms underlying paromomycin resistance and the high level of protein synthesis activity are apparently different.