Mutations in rsmG, Encoding a 16S rRNA Methyltransferase, Result in Low-Level Streptomycin Resistance and Antibiotic Overproduction in Streptomyces coelicolor A3(2)

Certain str mutations that confer high- or low-level streptomycinresistance result in the overproduction of antibiotics by Streptomycesspp. The str mutations that confer the high-level resistanceoccur within rpsL, which encodes the ribosomal protein S12,while those that cause low-level resistance are not as wellknown. We have used comparative genome sequencing to determinethat low-level resistance is caused by mutations of rsmG, whichencodes an S-adenosylmethionine (SAM)-dependent 16S rRNA methyltransferasecontaining a SAM binding motif. Deletion of rsmG from wild-typeStreptomyces coelicolor resulted in the acquisition of streptomycinresistance and the overproduction of the antibiotic actinorhodin.Introduction of wild-type rsmG into the deletion mutant completelyabrogated the effects of the rsmG deletion, confirming thatrsmG mutation underlies the observed phenotype. Consistent withearlier work using a spontaneous rsmG mutant, the strain carrying ${Delta}$ rsmG exhibited increased SAM synthetase activity, which mediatedthe overproduction of antibiotic. Moreover, high-performanceliquid chromatography analysis showed that the ${Delta}$ rsmG mutant lackeda 7-methylguanosine modification in the 16S rRNA (possibly atposition G518, which corresponds to G527 of Escherichia coli).Like certain rpsL mutants, the ${Delta}$ rsmG mutant exhibited enhancedprotein synthetic activity during the late growth phase. UnlikerpsL mutants, however, the ${Delta}$ rsmG mutant showed neither greaterstability of the 70S ribosomal complex nor increased expressionof ribosome recycling factor, suggesting that the mechanismunderlying increased protein synthesis differs in the rsmG andthe rpsL mutants. Finally, spontaneous rsmG mutations aroseat a 1,000-fold-higher frequency than rpsL mutations. Thesefindings provide new insight into the role of rRNA modificationin activating secondary metabolism in Streptomyces.

INTRODUCTION

Top

INTRODUCTION

One of the most intriguing challenges in biology is the elucidationof the mechanisms whereby cells switch from primary to secondarymetabolism in response to extracellular nutritional conditions.Among prokaryotes, Streptomyces spp. provide a tractable experimentalsystem for studying such mechanisms due to their display ofa wide range of adaptations to extreme nutrient limitations,including the production and secretion of antibiotics and enzymesand the formation of aerial mycelium and spores (4, 9). Streptomycescoelicolor A3(2) is the best-characterized strain in this genusand has been used to study mechanisms regulating antibioticproduction (7). This strain produces several chemically diverseantibiotics, including the red-pigmented tripyrrole undecylprodigiosin(Red) and the deep blue-pigmented polyketide actinorhodin (Act).We previously reported that the introduction of certain strmutations that confer streptomycin (Sm) resistance enhancesAct production in both S. coelicolor A3(2) and S. lividans (12,35). Moreover, an impaired ability to produce Act, resultingfrom certain developmental mutations (relA, relC, and brgA)in S. coelicolor, can be circumvented by introducing str mutations(21, 23, 35). These str mutants have been classified into twotypes. Type I mutants carry a mutation within rpsL, which encodesthe ribosomal protein S12, and exhibit high-level Sm resistance(MIC, >100 µg/ml), whereas type II mutants possessthe wild-type rpsL gene and exhibit low-level Sm resistance(MIC, 5 to 10 µg/ml).

We previously demonstrated that rpsL mutant ribosomes carryingthe K88E substitution in S12 (and K88R mutant ribosomes in Streptomycesalbus) are more stable than wild-type ribosomes at low magnesiumconcentrations, indicating that this increase in stability couldenhance protein synthesis (27, 38). We later found that increasedexpression of the translation factor ribosome recycling factor(RRF) also contributes to the enhanced protein synthesis observedduring the late growth phase in the K88E rpsL mutant. This ledus to conclude that both the greater stability of the 70S ribosomeand the elevated levels of RRF caused by the K88E rpsL mutationare responsible for the enhanced protein synthesis seen duringthe late growth phase and that this underlies the observed overproductionof antibiotic in the K88E rpsL mutant (13).

In contrast to the pivotal role played by changes in ribosomalfunction in type I (rpsL) mutants, we found that a type II mutant,strain KO-179 (str-19), displays markedly enhanced expressionof S-adenosylmethionine (SAM) synthetase (25). The importanceof SAM synthetase activity in initiating Act production is supportedby the results of RNase protection assays, which showed thatoverexpression of metK (encoding SAM synthetase) stimulatesthe expression of a positive regulatory gene (actII-ORF4) forthe act gene cluster, and by the finding that the exogenousaddition of SAM to the culture medium induces Act biosynthesisin wild-type cells (25). Similar results were reported for S.lividans (19). The molecular mechanism underlying the overexpressionof SAM synthetase in type II str mutants, however, remains unknown.

Our ultimate aim is to develop "ribosome engineering" (22) asa rational approach to taking full advantage of bacterial capabilities.Toward that end, a detailed understanding of the mechanism(s)underlying the processes outlined above will increase our understandingof how enhanced production of Act occurs in str mutants. Smwas first shown to be a particularly potent drug against Mycobacteriumtuberculosis in 1944 (34), and mutants resistant to Sm werereported as early as 1946 (20). However, the mechanism underlyinglow-level resistance to Sm (i.e., the type II mutation) hasremained obscure. In the present work, we successfully identifieda previously unknown mutation within rsmG (the gene encodinga 16S rRNA methyltransferase) that confers low-level Sm resistance.Moreover, further analysis of this mutation provided new insightinto the role of rRNA modification in activating secondary metabolismin Streptomyces.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Bacterial strains and culture conditions. Spores from S. coelicolor A3(2) strain 1147 (a prototrophicwild-type strain) and its derivatives were harvested from culturesgrown on GYM agar medium (35). Escherichia coli strains DH5 ${alpha}$ and GM2163 were used for routine DNA manipulation. R4 (35),R4C (R4 supplemented with 0.2% Casamino Acids), R5^– (14),and R5^– media without sucrose (R5MS) were used for Actproduction, with the latter assayed as described previously(18). Yeast extract malt extract (YEME) medium (18) and R5^–medium were used for studying in vitro protein synthesis andSAM synthetase activity, respectively. When necessary, agarmedium was inoculated with about 10⁶ spores of S. coelicolor,the plates were covered with cellophane, and the cultures weregrown at 30°C. Spontaneous low-level Sm-resistant mutants(rsmG mutants) were generated from wild-type strain 1147 (MIC,1 µg/ml) on GYM agar plates containing 3 µg/ml Sm.Mutants exhibiting high-level Sm resistance (rpsL mutants) wereobtained by plating spores on GYM agar medium containing 100µg/ml Sm. Serial dilutions of the cell suspension werealso plated on media without Sm to determine the number of viablecells in the original suspension. To measure the frequency ofresistant mutants, single colonies were isolated, and cellsoriginating from each of about 10 to 20 clones were examinedseparately.

Isolation and manipulation of DNA and RT-PCR analysis. Plasmid isolation, restriction enzyme digestion, ligation, andtransformation of E. coli and Streptomyces were performed asdescribed previously (18, 33). PCR was performed using TaKaRaLA Taq or TaKaRa Ex Taq enzyme with GC buffer I (TaKaRa, Tokyo).Unless stated otherwise, strain 1147 genomic DNA was used asthe PCR template. Construction of the plasmid pXEmetK, containinga metK-xylE transcription fusion element, and reverse transcription-PCR(RT-PCR) analysis were performed as described previously (25).

Deletion of rsmG. A 1-kb region 5' of rsmG was amplified by PCR using primersrsmG-F (5'-ACGAGAATTCGCGCAAGTTCATCAACGGTCTG-3') and rsmG-R (5'-CTGCGGATCCCACGGGGACCGTCCTTCCGTATG-3'),containing EcoRI and BamHI sites (underlined), respectively.A second 1-kb region 3' of rsmG was amplified using primersrsmG-F2 (5'-GACGGGATCCATAGCGTCCGGCTTCGCTCGGAT-3') and rsmG-R2(5'-TCGGAAGCTTACCAGCACGTCGTAGATGGAAG-3'), containing BamHI andHindIII sites (underlined), respectively. The PCR fragmentswere inserted between the EcoRI and HindIII sites of pK19mob,using three-fragment ligation, and the rsmG in-frame deletionconstruct was shuttled into pGM160::oriT (provided by HaruoIkeda, Kitasato University) to yield pGMDrsmG. E. coli ET12567/pUZ8002was transformed with the recombinant plasmid and mated intoS. coelicolor 1147. Transformants were selected with thiostreptonand purified by streaking onto thiostrepton-containing plates.To obtain single crossover recombinants, purified transformantswere cultured on thiostrepton-containing plates at 37°C(pGM160::oriT carries a temperature-sensitive replicon derivedfrom pSG5 and cannot replicate at 37°C). Thiostrepton-resistantsingle-crossover recombinants were subcultured by two roundsof streaking in the absence of thiostrepton at 37°C to obtaindouble-crossover recombinants in which the delivery plasmidwas lost from the cells. Serial dilutions of the resulting sporeswere plated, and the resulting colonies were tested for thiostreptonsensitivity. Colonies with a thiostrepton-sensitive phenotypewere selected, and the correct deletion of rsmG was confirmedby PCR using gene-specific primers and DNA sequencing. The ${Delta}$ rsmGmutant strain KO-656 was used for further analysis.

Complementation of the ${Delta}$ rsmG mutation. The S. coelicolor rsmG gene was PCR amplified using primersrsmG-F and rsmG-R2, as described above, and the resulting DNAfragment was cloned between the EcoRI and HindIII sites of the ${Phi}$ C31-derived integrating plasmid pTYM18 (28) (provided by HiroyasuOnaka, Toyama Prefectural University) to yield the plasmid pTYM-rsmG.E. coli ET12567/pUZ8002 cells were transformed with this plasmidand mated into the S. coelicolor ${Delta}$ rsmG strain. Transformantswere then selected with thiostrepton and used for phenotypicanalysis.

Analysis of in vitro methylation profiles of 16S rRNA. 16S rRNAs were extracted from 30S subunits isolated from S.coelicolor wild-type and ${Delta}$ rsmG strains. An aliquot (25 µg)of each extract was digested for 3 h at 37°C with nucleaseP1 (3 U) and alkaline phosphatase (0.04 U) in a 25-µlreaction mixture containing 20 mM HEPES-KOH (pH 7.5). The resultingnucleosides were analyzed by high-performance liquid chromatography(HPLC) using an Inertsil ODS-3 column (250 by 2.1 mm; GL Science,Japan) as described previously (16), except that the pH of solventA was adjusted to 4.0 instead of 5.3.

Enzyme assays and in vitro protein synthesis. SAM synthetase activity was measured as described previously(25), except that the cells were grown on R5^– agar. Cell-freetranslation of green fluorescent protein (GFP) mRNA and preparationof the S-150 fraction and ribosomes were performed as describedpreviously (13). Western blotting analysis of the RRF was alsoperformed as described previously (13). For sucrose gradientsedimentation analysis, ribosomal 70S complexes were sedimentedusing a Biocomp Piston Gradient Fractionator (Towa Kagaku) equippedwith an Atto Bio-Mini UV monitor (27).

RESULTS

Top

RESULTS

Identification of a str-1 mutation conferring low-level Sm resistance. Because of the inability of conventional genetic strategiesto identify mutations conferring low-level Sm resistance, weutilized comparative genome sequencing (CGS), a new method (1)that uses microarray-based DNA sequencing to identify singlenucleotide polymorphisms (SNPs) and insertion-deletion siteswithin the genome. The first phase of CGS (mapping) consistsof a genome-wide mutation analysis entailing hybridization ofmutant and reference strain genomic DNA to an array containingthe M145 reference genome with 7- to 8-bp probe spacing. Weutilized the mutant KO-132 (relA str-1) strain, showing low-levelSm resistance due to a str-1 mutation, as the source of mutantgenomic DNA, and the M145 strain as the source of referencegenomic DNA. Because the str-1 mutation had previously beenmapped to the 7-o'clock position on the S. coelicolor chromosome(21), CGS was conducted for 1.2 Mbp around this region. We identifieda putative SNP within the gene SCO3885 (Fig. 1), which was confirmedby direct sequencing to be a deletion mutation (deletion of488A). The homolog of this gene in E. coli was described asgidB based on its relation to glucose-mediated inhibition ofcell division (39), but we recently renamed it rRNA small subunitmethyltransferase gene G (rsmG) (26). Strikingly, six otherstr mutants (Table 1, KO-133 to KO-179) showing low-level Smresistance were all found to carry a mutation within the rsmGgene, strongly suggesting that rsmG mutations are responsiblefor this Sm resistance. These mutations were characterized bythe frequent appearance of insertion mutations that resultedin stop codons just downstream of the mutations.

View larger version (28K):
[in this window]
[in a new window]

FIG. 1. SignalMap (NimbleGen) representation of CGS analysis of the str-1 mutant strain KO-132. The lowest two traces show the signal intensities for the M145 wild-type (green) and str-1 mutant (blue) hybridizations; the red trace above shows their ratio. The top line depicts an SNP confirmed by sequencing.

View this table:
[in this window]
[in a new window]

TABLE 1. Location of mutation in the rsmG gene and resulting amino acid exchange in RsmG

Deletion of the rsmG gene. The str-1 mutation, as well as other str mutations (Table 1),was found to activate Act production when it was introducedinto the wild-type and the relA and brgA mutant strains, whichare otherwise Act deficient (35). To clearly demonstrate thatthe rsmG mutation is responsible for the observed phenotype(i.e., increased resistance to Sm and a greater ability to produceantibiotic), we performed a set of gene disruption experiments.The constructed rsmG deletion mutant (strain KO-656 carrying ${Delta}$ rsmG) showed increased resistance to Sm, as did the rsmG (str-19)mutant KO-179 strain (Fig. 2), and both showed markedly increasedAct production both on plates and in liquid cultures (Fig. 3A and B),strongly suggesting that these rsmG mutations are responsiblefor the observed phenotypes. Moreover, the introduction of aplasmid containing wild-type rsmG into ${Delta}$ rsmG cells completelyeliminated resistance to Sm and abrogated the positive effectof the ${Delta}$ rsmG mutation on antibiotic production (data not shown).The observed positive effect of the ${Delta}$ rsmG mutation on Act productionwas ascribed to transcriptional activation of the pathway-specificpositive regulatory gene actII-ORF4 (Fig. 3C).

View larger version (27K):
[in this window]
[in a new window]

FIG. 2. Effect of deleting the rsmG gene on the level of Sm resistance. The 1147 (wild-type [WT]) and KO-656 ( ${Delta}$ rsmG) strains were grown for 2 days on GYM agar with (+SM) or without (–SM) 2 µg/ml Sm. Strain KO-179 (rsmG mutant [str-19]) served as a reference.

View larger version (54K):
[in this window]
[in a new window]

FIG. 3. Effect of deleting the rsmG gene on Act production and expression of actII-ORF4. (A) Act production by S. coelicolor strains 1147 (wild-type [WT]) and KO-656 ( ${Delta}$ rsmG). Upper row shows the reverse side of the plates. Strains were incubated on an R4C agar plate for 4 days. (B) Act production by S. coelicolor strains 1147 and KO-656 in R5^– liquid medium. (C) Expression of actII-ORF4 mRNA by cells grown on R4C agar plates for the indicated times, as determined by RT-PCR. (D) Effect of rsmG mutation on Act and Red production by S. lividans grown on R4 agar plates for 6 days. The upper (left) and reverse (right) sides of the plates are shown. Blue and red represent actinorhodin (Act) and undecylprodigiosin (Red), respectively.

Drug-resistant bacteria are generally believed to grow moreslowly than susceptible bacteria because mutations that conferresistance also reduce the organism's overall fitness, a phenomenonknown as "cost of resistance" (2). However, the S. coelicolor ${Delta}$ rsmG mutant grew as well as the parent strain in YEME (Fig.6A), GYM, and R4C media (data not shown).

View larger version (12K):
[in this window]
[in a new window]

FIG. 6. Profiles of growth and in vitro protein synthesis in wild-type 1147 (WT) and ${Delta}$ rsmG (KO-656) strains. (A) Growth in YEME medium at 30°C was monitored by measuring the optical density at 450 nm (OD₄₅₀). The zero time point represents 22 to 24 h after inoculation of fresh spores, when the OD₄₅₀ was 0.2; "S2" indicates the early stationary phase (see reference 13). (B) In vitro synthesis of GFP using wild-type and ${Delta}$ rsmG ribosomes prepared from cells grown to S₂ phase. Strain KO-178 (K88E rpsL mutant) served as a reference strain. Equal aliquots (10 µl) of reaction mixture were withdrawn at the indicated times and subjected to electrophoresis in 10% polyacrylamide gels. The intensity of the GFP bands was determined by scanning the fluorographs. (C) Effects on GFP synthesis of cross-mixing the S-150 fractions and ribosomes from wild-type and mutant cells grown in YEME medium to stationary (S₂) phase. Cell-free translation of GFP mRNA was performed as described in the panel B legend. (Upper panel) Fluorographs of synthesized GFP. (Lower panel) Relative levels of GFP synthesis. (D) Expression profile of RRF protein in wild-type (1147) and mutant ( ${Delta}$ rsmG) strains. Strain KO-178 (K88E rpsL mutant) was the reference strain.

RsmG functions as a ribosome modulator by targeting 16S rRNA. RsmG protein has a putative SAM binding motif within its primarystructure (17) (see Fig. S1 in the supplemental material), andthe crystal structure of E. coli RsmG shows that this proteincontains a SAM-dependent methyltransferase fold within its ternarystructure (32). We recently showed that E. coli RsmG catalyzesa SAM-dependent 7-methylguanosine (m⁷G) modification of G527in the 16S rRNA (26). By analyzing the nucleosides using reversed-phaseHPLC, we confirmed that S. coelicolor RsmG also catalyzes anm⁷G modification within the S. coelicolor 16S rRNA (Fig. 4).As expected, the peak corresponding to m⁷G was absent from the ${Delta}$ rsmG mutant, demonstrating that RsmG is responsible for thein vivo methylation of 16S rRNA. Since the m⁷G modificationis restricted to a single position in E. coli 16S rRNA and RsmGhas equivalent functions in E. coli and S. coelicolor (26),these findings suggest that RsmG modifies G518 in S. coelicolor,corresponding to G527 in E. coli.

View larger version (16K):
[in this window]
[in a new window]

FIG. 4. HPLC profile of 16S rRNA nucleosides from wild-type (upper panel) and ${Delta}$ rsmG mutant (lower panel) strains. 16S rRNA was isolated, digested completely with nuclease P1 and alkaline phosphatase, and analyzed by HPLC. The peak position for m⁷G was determined using standard m⁷G.

The rsmG mutant exhibits increased SAM synthetase activity. We previously reported that S. coelicolor strain KO-179 (str-19),which harbors a Val 10 $->$ Cys mutation within rsmG (Table 1), exhibitsenhanced expression of SAM synthetase, eventually leading tothe overproduction of Act (25). This was confirmed by usingthe ${Delta}$ rsmG mutant, in that measurements of enzyme activity revealed,as expected, that the ${Delta}$ rsmG mutant displayed a 5- to 10-foldincrease in SAM synthetase activity during the late growth phase(Fig. 5A). Furthermore, dramatically enhanced transcriptionof metK (a gene encoding SAM synthetase) accounted for the elevatedSAM synthetase activity (Fig. 5B), indicating that by actingat the level of transcription, the rsmG mutation has a positiveeffect on MetK expression.

View larger version (12K):
[in this window]
[in a new window]

FIG. 5. Effect of deleting rsmG on SAM synthetase activity. (A) SAM synthetase activity in wild-type (1147) and ${Delta}$ rsmG mutant (KO-656) strains. Cells were grown on R5^– agar covered with cellophane. Samples were taken at the indicated times, and SAM synthetase activity was determined. One unit of activity is defined as the amount of enzyme that changed the optical density at 340 nm at a rate of 12.4/min. (B) Expression of metK-xylE fusion element, as determined by quantitative catechol dioxygenase assays in wild-type and ${Delta}$ rsmG mutant strains grown on R5^– agar medium. XylE activity was determined as described previously (25).

The ${Delta}$ rsmG mutant ribosome exhibits enhanced translational activity. Certain rpsL mutations (e.g., the K88E substitution in S12)lead to enhanced protein synthetic activity during the lategrowth phase, in addition to high-level Sm resistance (13, 27).To test whether rsmG mutations also confer this phenotype, wemeasured the in vitro translational activity of ribosomes isolatedfrom wild-type and ${Delta}$ rsmG cells. Washed ribosomes and the S-150fraction were prepared from cells grown to early (S₂) stationaryphase in YEME medium (Fig. 6A) and were used to assemble thein vitro translation system, using GFP mRNA as a template. Inwild-type extracts, the rate of GFP synthesis was maximal duringthe mid-exponential phase but declined during the stationaryphase (13). By contrast, ribosomes isolated from stationary-phase ${Delta}$ rsmG cells exhibited a high-level of GFP synthetic activity,as did ribosomes from K88E rpsL mutant cells (Fig. 6B); ribosomesprepared from ${Delta}$ rsmG cells were 4.5-fold more active than thoseobtained from wild-type strain 1147 during the stationary phase.Similar results were observed when the activity was comparedwith cells grown to mid (S₃)-stationary phase (data not shown).Enhanced translational activity during the stationary phasethus appears to be a characteristic property of the ${Delta}$ rsmG mutant.

In the K88E rpsL mutant, the enhanced protein synthesis activityduring the late growth phase was found to reflect the increasedstability of the 70S ribosomal complex and the higher levelsof RRF in the S-150 fraction (13, 27). Our cross-mixing experiments,however, showed that the S-150 fraction prepared from ${Delta}$ rsmG cellsdid not contribute to the enhancement observed in GFP synthesis(Fig. 6C). Consistent with that result, Western analysis revealedthat, unlike that in the K88E rpsL mutant, the S-150 fractionin the ${Delta}$ rsmG mutant did not show increased RRF levels (Fig. 6D).Moreover, the ${Delta}$ rsmG mutation did not lead to increased stabilityof the 70S ribosomal complex at low (1 mM) Mg²⁺ concentration.By contrast, 70S ribosomal complexes from the K88E rpsL mutantshowed greater stability than those from wild-type cells orthe ${Delta}$ rsmG mutant (see Fig. S2 in the supplemental material).These findings indicate that although the rpsL and rsmG mutantstrains both exhibit enhanced protein synthesis activity duringlate growth phase, the mechanisms underlying this enhanced activityare largely different. Although ribosomes from the ${Delta}$ rsmG mutantshowed enhanced protein synthesis activity, ribosomes from thewild-type strain, into which a high-copy-number plasmid containingthe metK gene had been introduced (25), did not (data not shown),indicating that increased SAM synthetase activity did not causethe enhanced protein synthesis activity.

Characterization of the rsmG mutation. In bacteria, spontaneous mutations leading to high-level Smresistance (e.g., a 100-fold increase in MIC) generally emergeat a low frequency, with the majority arising within rpsL (10).For example, in S. coelicolor wild-type strain 1147 (MIC, 1µg/ml), the frequency of emergence of spontaneous mutantswith high-level Sm resistance (MIC, $≥$ 100 µg/ml) was aslow as 10^–10 to 3 x 10^–10. By contrast, mutantswith low-level Sm resistance (MIC, 5 to 10 µg/ml) emergedat much higher frequencies, ranging from 10^–6 to 8 x 10^–6.Gene sequencing showed that all 18 low-level Sm-resistant mutantsharbored a point, deletion, or insertion mutation within rsmGand that all produced much more Act on R4C agar medium thandid the parental strain 1147 (several isolates [e.g., KO-660to KO-666] are listed in Table 1). Likewise, rsmG mutants ofS. lividans wild-type strain 1326 (Table 1) produced Act onR4 agar, although the parental organism did not (for example,the rsmG mutant strain KO-690 [21C $->$ CC, frameshift mutation, Cinsertion at position 21] in Fig. 3D). Together with the ${Delta}$ rsmGmutant, all of these rsmG mutants of S. coelicolor and S. lividanshad reduced abilities to form aerial mycelium (and thus, sporulation).

We found that the ${Delta}$ rsmG strain produced mutants showing resistanceto high-level Sm (100 µg/ml) at frequencies ranging from10^–8 to 3 x 10^–7, 100- to 1,000-fold higher thanthe frequencies observed with the wild-type strain 1147. Most(14/20) of these highly resistant ${Delta}$ rsmG mutants had rpsL mutations,including the previously unreported T41I mutant. Moreover, certainrsmG rpsL double mutants (e.g., ${Delta}$ rsmG rpsL [K88E] and ${Delta}$ rsmG rpsL[K88R]) displayed a greater ability to produce antibiotics thansingle mutants did (see Table S1 in the supplemental material).

DISCUSSION

Top

DISCUSSION

The str-19 and str-1 mutants have been isolated from S. coelicolorwild-type strain 1147 and from the relA mutant strain M570,respectively, based on their low-level Sm resistance, accompaniedby markedly increased Act production in the former and completerestoration of impaired Act production in the latter (21, 35).Further analysis of these str mutants showed that enhanced expressionof SAM synthetase caused by the mutation was responsible forthe overproduction of Act (25). However, the precise locationof the str mutations conferring low-level Sm resistance wasunknown. In the present work, we used the CGS technique to successfullyshow that this mutation lies within rsmG and then characterizedthe cellular metabolic changes caused by this mutation.

We demonstrated unambiguously that the loss of the m⁷G modificationwithin the 16S rRNA results in resistance to Sm, thereby providinga molecular basis for rsmG mutation-induced Sm resistance. Thesite of the methylation is highly likely to be G518, which correspondsto G527 of E. coli, as the RsmG proteins of S. coelicolor andE. coli appear to be functionally equivalent and, thus, areinvolved in the same biochemical process (26). It is noteworthythat G527 is located within the 530 loop of 16S rRNA, whichappears to play a key role in mediating the accuracy of proteinsynthesis, and that this invariant nucleotide directly interactswith Sm (5, 8, 24, 30, 31, 37). Consequently, our finding thatthe failure to methylate G527 residue leads to Sm resistanceis striking.

Homologues of rsmG are highly conserved among eubacteria, soit was somewhat surprising that despite the apparently importantcontribution made by RsmG to ribosomal function, the deletionof rsmG had no effect on the growth of either S. coelicolor(in this study) or E. coli (26). The high frequency of emergenceof spontaneous rsmG mutants is likely due to the dispensabilityof this gene, allowing cells to remain viable. It is also noteworthythat mutants showing high-level Sm resistance (i.e., rpsL mutants)arose at much higher frequencies in rsmG mutants than in wild-typecells. We do not know at present the mechanism underlying thehigh-frequency emergence of high-level Sm resistance, but itis not likely that RsmG functions as an antimutator-like protein,since rsmG mutation did not affect the frequency at which mutantsresistant to antibiotics other than Sm emerged (26). Nonetheless,our finding that rsmG rpsL double mutants have a greater abilityto produce antibiotic (see Table S1 in the supplemental material)is intriguing in considering strategies for strain improvement.

We observed a causal relationship between rsmG mutation andupregulation of metK expression (Fig. 5), which together withearlier work (15, 19, 25) establishes the significance of SAMsynthetase activity in initiating antibiotic production in S.coelicolor A3(2) and other Streptomyces spp. As shown by metK-xylEfusion analysis, enhanced expression of MetK protein in the ${Delta}$ rsmG disruptant is apparently achieved through the upregulationof transcription (Fig. 5B). It is unclear, however, how thersmG mutation dramatically upregulates metK transcription.

The principal regulator of Act production in S. coelicolor appearsto be the availability of the pathway-specific transcriptionalregulatory protein ActII-ORF4, a threshold concentration ofwhich is required for efficient transcription of its cognatebiosynthetic structural genes (3, 11). It is noteworthy thatthe rsmG mutation enhances actII-ORF4 transcription (Fig. 3C),as it suggests SAM-dependent protein methylation may be involvedin controlling the activity of the regulatory proteins encodedby such developmental genes. Alternatively, DNA or RNA methylationmay be involved in the expression of these regulatory genes,or SAM itself may be an inducer of Act synthesis. In relationto this notion, recent findings suggest that exogenous additionof SAM enhances the expression of BldK, an oligopeptide transporterimportant in the regulation of S. coelicolor differentiation,as well as the transcription of the global regulatory S. griseusgenes adpA and strR, leading to Sm overproduction (29, 36).

We found that both the ${Delta}$ rsmG and K88E rpsL mutants exhibitedenhanced protein synthetic activity during the late growth phase(Fig. 6), which is consistent with our hypothesis (13) thatthe capacity of a cell to synthesize protein during late growthphase is indicative of its ability to accelerate the onset ofsecondary metabolism and to produce biosynthetic enzymes. Althoughwe do not yet know how these mutations mediate preferentialgene transcription, it is conceivable that the expression ofpathway-specific regulatory genes is governed by higher-orderregulatory proteins and that expression of the latter presumptiveregulatory proteins may be significantly affected under conditionsassociated with enhanced protein synthesis during the stationaryphase in the mutants. In that regard, it is noteworthy thatthe 70S ribosome in the ${Delta}$ rsmG mutant was not more stable (seeFig. S2 in the supplemental material) and that this mutant didnot show increased levels of RRF in the S-150 fraction, allof which would contribute to protein synthetic activity. Apparently,the enhanced protein synthesis observed in these different Sm-resistantmutants (rpsL versus rsmG) is mediated by largely differentmechanisms.

ACKNOWLEDGMENTS

This work was supported by grants to K.O. from the OrganizedResearch Combination System and the Effective Promotion of JointResearch of Special Coordination Funds (the Ministry of Education,Culture, Sports, Science and Technology of the Japanese Government).

We are grateful to Hiroyasu Onaka and Haruo Ikeda for theirgenerous gifts of plasmids used here and to the GenefrontierCorp. (Tokyo) for supporting the mutation search using the comparativegenome sequencing technique.

FOOTNOTES

* Corresponding author. Mailing address: National Food Research Institute, 2-1-12 Kannondai, Tsukuba, Ibaraki 305-8642, Japan. Phone: 81-29-838-8125. Fax: 81-29-838-7996. E-mail: kochi{at}affrc.go.jp

Published ahead of print on 23 March 2007.

Supplemental material for this article may be found at http://jb.asm.org/.

This paper is dedicated to Keith F. Chater upon his retirementfrom the John Innes Institute.

REFERENCES

Top