INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

One of the significant bacterial regulatory systems is the stringent response which is initiated by nutrient limitation and causes an immediate cessation of RNA accumulation and other cellular reactions. The guanine nucleotides ppGpp (guanosine 5'-diphosphate 3'-diphosphate) and pppGpp (guanosine 5'-triphosphate 3'-diphosphate) are believed to be responsible for this stringent response. Mutants with mutations in the relA or relC (=rplK) genes (which code for the ppGpp synthetase and the ribosomal L11 protein, respectively) fail to synthesize (p)ppGpp (for reviews, see references 3 and 14). Recently, through working with Streptomyces coelicolor A3(2), an unambiguous correlation was established between ppGpp synthesis and antibiotic production as demonstrated from the results of three independent laboratories using various approaches at the molecular level (4, 28, 35). We have found (35, 43) that the impairment in antibiotic production resulting from a relA or relC mutation in S. coelicolor A3(2) could be completely restored by introducing mutations conferring resistance to streptomycin (str). These mutations result in the alteration of the Lys-88 amino acid in ribosomal protein S12 (rpsL gene product) to Glu or Arg (35, 43). No accompanying restoration of ppGpp synthesis was detected in these relA str or relC str mutants. It is therefore apparent that acquisition of certain str mutations allows antibiotic production to be initiated without the requirement for ppGpp. This offers a possible strategy for improving the antibiotic productivity of wild-type prokaryotic microorganisms.

The initiation of antibiotic biosynthesis (so-called secondary metabolism) usually occurs at the transition between vegetative growth and morphological development, such as the sporulation stage of the organism, and considerable effort has been directed towards gaining a detailed understanding of the mechanism involved (reviewed in references 5-7, 22, 26, 38, and 58). Members of the genera Streptomyces, Bacillus, and Pseudomonas are soil bacteria that produce a high proportion of agriculturally and medically important antibiotics. The development of rational approaches to improve the production of antibiotics from these organisms is therefore of considerable industrial and economic importance. This paper describes the effect of the introduction of str mutations on antibiotic production in these organisms.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and preparations of mutants. The strains of Streptomyces spp. and other bacteria used in this study are listed in Table 1. Bacillus subtilis 168 is a standard strain frequently used for studying sporulation. Strains 60009, 61884, 61883, and 61953 are derivatives of strain 168. The spontaneous streptomycin-resistant mutants of each organism were obtained as colonies that grew within 7 to 14 days (for Streptomyces spp.) or within 3 to 7 days (for other bacteria) after spores or cells were spread on glucose-yeast extract-malt extract (GYM) agar containing streptomycin at a stated concentration (see below). The mutants obtained were used for subsequent study after single-colony isolation was done.

Media and growth conditions. GYM medium, nutrient sporulation medium, and synthetic medium (S7 medium) were described previously (32, 33). NG medium, a medium developed for antibiotic production by Bacillus subtilis, contained (per liter) 10 g of nutrient broth (Difco), 10 g of glucose, 2 g of NaCl, 5 mg of CuSO₄ · 5H₂O, 7.5 mg of FeSO₄ · 7H₂O, 3.6 mg of MnSO₄ · 5H₂O, 15 mg of CaCl₂ · 2H₂O, and 9 mg of ZnSO₄ · 7H₂O (adjusted to pH 7.2 with NaOH). All strains were stocked after growing on GYM agar at 30°C.

Production and determination of antibiotics. All strains, including B. subtilis, were incubated at 30°C with shaking. The culture conditions for the production of each antibiotic were optimized as described below.

(i) Fredericamycin. Streptomyces chattanoogensis ISP 5002 was precultured in galactose-glycerol-corn steep liquor medium for 2 days. Then, cells were inoculated into the production medium followed by 4 days of cultivation. The production medium consisted of (per liter) 5 g of L-phenylalanine, 1.5 g of ammonium sulfate, 0.5 g of K₂HPO₄, 0.5 g of KH₂PO₄, and 2 g of CaCO₃ (adjusted to pH 7.2 with NaOH). Concentration of fredericamycin produced was determined by high-performance liquid chromatography analysis.

(ii) Actinomycin. Streptomyces antibioticus 3720 (= ATCC 14888) was grown in peptone (NZ-amine; type A) medium for 30 h. Cells were inoculated into a production medium (galactose-glucose-glutamate medium) and cultivated for 4 days as described previously (31). The concentration of actinomycin produced was determined spectrometrically at 443 nm after extraction with ethyl acetate.

(iii) Formycin. Streptomyces lavendulae MA406 A-1 was cultivated in maltose-polypeptone-yeast extract (MPY) medium for 2 days as described previously (30). Concentrations of formycin were determined by bioassay using Xanthomonas oryzae as a test organism.

(iv) Pyrrolnitrin. Pseudomonas pyrrocinia 2327 (=ATCC 15958) was precultured in glucose-polypeptone-meat extract medium (2) for 2 days. Cells (0.1 ml) were then inoculated into the same medium (5 ml) and cultivated for 3 days. Concentrations of pyrrolnitrin were determined by a bioassay using Candida albicans as a test organism.

(v) FR900493. Bacillus cereus 2045 (=FERM BP-1791) was precultured in bouillon medium for 10 h. Cells (0.1 ml) were inoculated into 5 ml of production medium consisting of (per liter) 20 g of polypeptone, 20 g of corn steep liquor, and 5 g of NaCl (adjusted to pH 7.5 with NaOH). Concentrations of FR900493 {3-[(5-aminomethyl-3,4-dihydroxytetrahydrofuran-2-yl)oxy]-2-[(3-aminopropyl) methylamino]-3-[5-(3,4-dihydro-2,4-dioxo-2H-pyrimidin-1-yl)-3,4-dihydroxytet- rahydrofuran-2-yl]propionic acid} were determined by a bioassay using Staphylococcus aureus 209P.

(vi) B. subtilis antibiotic. Strains of B. subtilis were grown in NG medium (see above) plus requirements (Trp, Thr, or Asp [each at 50 µg/ml]) by direct inoculation of spores into the medium. Cultivation was carried out at 30°C for 30 h. Antibiotic production was determined by a bioassay using S. aureus 209P as described previously (34). Antibiotic activity is expressed as units per milliliter. One unit per milliliter produced a 1.5-mm-wide halo (diameter of clear area, 11 mm). Test solutions were diluted appropriately to produce haloes of this diameter.

Sporulation conditions for B. subtilis. The cells were inoculated at an optical density at 600 nm (OD₆₀₀) of 0.01 to 0.02 into flasks containing sporulation medium (33) plus requirements (Trp, Thr, or Asp [each at 50 µg/ml]) and were incubated at 37°C for the indicated times with vigorous shaking. The spore titer was measured by first heating the culture for 20 min at 75°C and then plating after appropriate dilution.

Nutritional shift-down and assay of ppGpp. B. subtilis cells were grown in 20 ml of synthetic medium (33) plus 1% (wt/vol) vitamin-free Casamino Acids (Difco) and requirements (Trp, Thr, or Asp [each at 50 µg/ml]). When cells had grown to mid-exponential phase (OD₆₀₀ = 2 to 2.5), they were collected on a membrane filter (0.45-µm pore size), immediately washed with 10 ml of synthetic medium, and the filters with cells were quickly transferred to 20 ml of the synthetic medium without Casamino Acids. After incubation for 15 min, cells were rapidly collected as just described, and the filters were then laid upside down on 10 ml of formic acid (1 M) in a plastic petri dish. After a 60-min incubation at 4°C, cells were removed by centrifugation. The supernatant was filtered through a syringe and vacuum evaporated. The residue was dissolved in 200 µl of water and used for the determination of ppGpp.

Detection of the rpsL gene mutation in B. subtilis. The rpsL gene of B. subtilis (as shown in the sequence obtained by H. Yoshikawa [DDBJ accession no. D64127]) was amplified from B. subtilis genomic DNA by PCR using synthetic oligonucleotide primers P1 (5'-CCACCTGGGTATGTGGGTT-3') and P2 (5'-GCATCTGAATATCCTCCCT-3'). The PCR products were directly sequenced manually with -³⁵S-dCTP or automatically by using the ABI PRISM 310 Genetic Analyzer. The chosen primers amplify the whole rpsL gene so that we can evaluate whether not finding a mutation in a particular gene indicates that the defect must be in another gene.

Replacement of the rpsL gene. Mutant rpsL genes were amplified by PCR with mutant genomic DNA (str-5, str-9, or str-10) as a template. The primers used were P3 (5'-CCTTGAATTCGCCTACAATTAATCAGCTAA-3', corresponding to positions 3 to 22 of the rpsL open reading frame [underlining indicates the cleavage site for EcoRI]) and P4 (5'-CCTTAAGCTTGCATCTGAATATCCTCCCT-3', corresponding to positions 27 to 47 of the gene outside the C-terminal end of the rpsL open reading frame [underlining indicates the cleavage site for HindIII). The resulting 479 bp of PCR product were digested with EcoRI and HindIII and then cloned into the corresponding sites of pAG58 (25). pAG58 is an expression vector which can integrate into the B. subtilis chromosome if the plasmid contains a DNA fragment homologous to part of the chromosome. This plasmid contains the chloramphenicol resistance gene (cam) as a selective marker in B. subtilis. The resulting plasmids, named pSR1, pSR150, and pSR87, were used to transform wild-type B. subtilis 168, and the transformants were first selected for streptomycin resistance. The concentration of streptomycin used for the selection of transformants was 100 µg/ml. The resulting str transformants were then screened for chloramphenicol sensitivity to obtain the clones with the mutant alleles, in which the gene replacement occurred by a double-crossover event (see Fig. 2).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Effect of str mutation on Streptomyces spp. We first examined the effect of a str mutation on antibiotic production in three Streptomyces spp., S. chattanoogensis (which produces fredericamycin), S. antibioticus (which produces actinomycin), and S. lavendulae (which produces formycin). When the spores of these Streptomyces spp. were spread and incubated on GYM agar containing 5 or 30 µg of streptomycin per ml, streptomycin-resistant (str) mutants developed after 7 to 14 days at a frequency of 10⁶ to 10⁸. These spontaneous str mutants were characterized from a wide variety of colonies by size, morphology, and pigment formation. S. lavendulae and S. chattanoogensis are shown as examples (Fig. 1). We examined antibiotic production using each of 50 to 100 str mutants that were selected randomly. Strikingly, in S. chattanoogensis nearly half of the str mutants tested exhibited a significantly increased ability (greater than fivefold) to produce fredericamycin. The highest productivity detected was 26-fold higher than that of the wild-type strain (Table 2). The level of streptomycin resistance of the high-level-antibiotic-producing (high-producing) strains ranged from 50 to 200 µg/ml. Similarly, strains producing high levels of actinomycin and formycin could be detected at a relatively high frequency (3 to 4%) among str mutants of S. antibioticus and S. lavendulae, respectively (Table 2). Thus, like actinorhodin production by S. coelicolor A3(2) (19), introduction of mutations conferring resistance to streptomycin was shown to be effective for improving the antibiotic productivity of the Streptomyces spp. tested.

View larger version (112K): [in this window] [in a new window]   FIG. 1.   Spontaneously generated streptomycin-resistant colonies of S. lavendulae and S. chattanoogensis. Spores (ca. 109) of wild-type S. lavendulae and S. chattanoogensis were spread on GYM agar plates containing 5 or 30 µg of streptomycin per ml, respectively, and incubated at 30°C for 10 days.

Effect of str mutations on Bacillus and Pseudomonas. Introduction of the str mutation also improved antibiotic productivity of bacteria such as Bacillus spp. and Pseudomonas spp. (Table 2). In B. cereus (which produces the nucleoside antibiotic FR900493) and P. pyrrocinia (which produces pyrrolnitrin), the frequency of antibiotic overproducing strains among str mutants ranged from 7 to 30%.

Characterization of str mutants derived from B. subtilis. We focused on B. subtilis for further investigation, since this organism offers a feasible system for genetic analysis at the molecular level.

Gene replacement between str and str⁺ alleles. In order to clearly demonstrate that the str mutations discussed above are responsible for the observed phenotype, i.e., the increased antibiotic-producing ability, gene replacement experiments were performed (see Materials and Methods). The plasmids pSR1, pSR150, and pSR87, which contain the EcoRI-HindIII fragment of the mutant rpsL gene from either strain KO-267 (str-5), KO-271 (str-9), or KO-272 (str-10), respectively, was used to transform the parent strain 168. Our strategy for gene replacement of rpsL with mutant alleles by a double-crossover event is illustrated in Fig. 2. Streptomycin-resistant but chloramphenicol-sensitive recombinants all overproduced antibiotic as much as the original mutant strains. Gene replacement of the mutant rpsL genes was confirmed by directly sequencing the PCR products. Thus, we concluded that str-5, str-9, and str-10 mutations (altering Lys-56 to Arg, Thr, and Gln, respectively) are responsible for the observed increase in antibiotic productivity.

View larger version (26K): [in this window] [in a new window]   FIG. 2.   Replacement of the rpsL locus by double-crossover recombination between chromosome and pSR1 containing a mutated rpsL allele. rpsLstr indicates the mutated rpsL gene (Lys-56 to Arg). The shaded area represents the mutation site. cam and rpsG indicate the chloramphenicol resistance gene and the ribosomal protein S7-encoding gene, respectively.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Our principal finding in this study was that the introduction of a specified str mutation into any bacterial genera gives rise to a marked increase in antibiotic productivity, thus expanding the previous preliminary work with S. coelicolor A3(2) (19). This novel breeding approach not only results in yielding high-producing strains but also makes it possible to generate these overproducing strains at a surprisingly high frequency (3 to 50% in general). Transferring this trait into other procaryotic microorganisms should therefore offer a convenient and effective method for improving antibiotic productivity. Indeed, as reported previously (43), introduction of a certain str mutation (altering Lys-88 to Glu in ribosomal protein S12) was effective even for activating the silent gene(s) in Streptomyces lividans, which is involved in production of the blue-colored antibiotic actinorhodin in this organism.

Elucidating the mechanism of initiation of so-called secondary metabolism has been the subject of a large number of publications, but very few have focused on ribosomal function or on the translational machinery. It is important to point out that among the ribosome-binding drugs tested so far, only streptomycin has been shown to be effective for improving antibiotic productivity (reference 19 and this study). The action of streptomycin on bacterial ribosomes has been studied in great detail (reviewed by Wallace et al. [54] and Cundliffe [8]), and among the numerous effects attributed to this drug, the misreading of mRNA codons is the best known. Positions in S12 affected by amino acid alterations which have previously been reported to confer streptomycin resistance in E. coli and Mycobacterium tuberculosis are Lys-43, Lys-88, and Pro-91 (11, 12, 21, 29, 53, 56). In S. coelicolor the alteration in Arg-86 also confers resistance (albeit low-level resistance) to streptomycin (43). The alteration in Lys-56 (corresponding to Lys-43 in E. coli and S. coelicolor) identified in B. subtilis in the present study therefore corresponds to one of these recognized positions (Table 3). Unlike those in E. coli and M. tuberculosis, the str mutations detected within the rpsL gene in B. subtilis were all represented by changes at the Lys-56 position (Table 3). As reported previously (43), mutation at Lys-43 in S. coelicolor does not result in an increase in antibiotic productivity, although it elicits a high level of resistance to streptomycin. Of particular interest is the fact that antibiotic productivity of B. subtilis increased in a hierarchical manner depending upon which species of amino acid replaced Lys at position 56 (Table 3). The set of mutants described in this paper therefore may offer a feasible system for studying the regulation of bacterial secondary metabolism at the translational level, together with the bldA (encoding leucine-tRNA) mutant of S. coelicolor as studied by Guthrie and Chater (17). We have thus demonstrated that an altered ribosomal protein S12, resulting from specific mutations in the rpsL gene which confer high level resistance to streptomycin, elicits the ability to overproduce antibiotics not only in streptomycetes but also in members of the family Bacillaceae. Although DNA sequencing was not attempted in this study for the gram-negative bacterium P. pyrrocinia, it is possible that the pyrrolnitrin overproducing str mutants (Table 2) also have a mutation in the rpsL gene.

In B. subtilis, in addition to the relA and relC mutations, a number of other mutations affecting the ability to produce antibiotic have been reported. These include abrA (which may be same as rev-4) and abrB (which is probably the same as absA and absB) (summarized by Piggot and Hoch [39]). sco mutations (scoA, scoB, scoC, and scoD) affecting sporulation control give rise to various degrees of overproduction of alkaline proteases (9). Mutations abrA, abrB, absA, and absB were all originally found by their ability to resist an antibiotic produced by the wild-type sporulating B. subtilis strain 168 (16, 24, 52). Some of these mutations compensate for the detrimental pleiotropic effects resulting from a spo0A mutation by restoring antibiotic and protease production, competence for transformation, and polymyxin resistance (51). The abrB locus has been cloned and sequenced (37). Mutations in this gene affect the transcription of a variety of genes, including aprE (which encodes subtilisin) (10), tycA (which encodes tyrocidine synthetase) (13, 27, 40), and spo0E and spoVG (57). Recent studies have demonstrated that the abrB gene encodes a 10.5-kDa protein which functions as either an activator or a repressor of the expression of a variety of genes by binding to the promoter regions of these genes (46, 55; reviewed briefly by O'Reilly and Devine [36]). Thus, the abrB locus appears to be the major locus responsible for regulating transition stage gene expression. Interestingly, all of the abrB mutants studied so far have been found to have alterations (missing or changed in polarity) in one or more of several different 50S ribosomal proteins (42, 50, 51). However, the causal relationship between the abrB mutation and the observed alteration in ribosomal proteins still remains obscure. The mutation rev-4 (which may be the same as abrA) has been reported to restore the suppressed sporulation caused by other mutations in the RNA polymerase, ribosomal proteins, and the protein synthesis factor EF-G (41). The rev-4 mutation has been located in orfR of the spo0F region (49). It is possible that the streptomycin-resistant mutants identified in the present study that do not have a mutation in the S12 protein (e.g., KO-274  KO-278 [Table 3]) harbor a mutation in a ribosomal protein other than S12 (or in an rRNA component). This notion may be supported by the fact that those str mutants demonstrated an increased sensitivity to certain ribosome-binding antibiotics such as thiostrepton, and moreover str mutants (as examined with KO-274) revealed an impaired ability to accumulate ppGpp (see Results). In relation to this argument Staal and Hoch (44) reported a new class of str mutation, called strB, which confers conditional streptomycin resistance and is distinct from the classical strA (= rpsL) locus.

There are detailed studies from several laboratories that could shed light on the relationship of various phenotypes to ppGpp levels. On the basis of the isolation and analysis of relaxed (relC) mutants of several Streptomyces spp., ppGpp has been shown to play a central role in triggering the onset of antibiotic production (30-32). In agreement with our results, ppGpp accumulation was noted to coincide with transcription of the pathway-specific activation genes for undecylprodigiosin (redD) and actinorhodin (actII-ORF4) in S. coelicolor (45, 47, 48), and for bialaphos (brpA) in Streptomyces hygroscopicus (20). Antibiotic production by B. subtilis is also abolished by mutations such as relA and relC, which cause a deficiency in ppGpp synthetase and ribosomal L11 protein, respectively (34). As will be reported in detail elsewhere, acquisition of certain str mutations by B. subtilis relA and relC mutants (61883 and 61953, respectively) restores the antibiotic productivity lost in these mutants, without any accompanying ppGpp accumulation. The dependence of B. subtilis on ppGpp for the initiation of antibiotic production is therefore apparently bypassed by certain str mutations, as was the case in S. coelicolor A3(2) (35, 43). The molecular basis for the observed role of the altered ribosomal protein S12 (and another putative ribosomal protein) is unclear. It is, however, conceivable that the specified ribosomal mutations led to a change in ribosomal structure which gave rise to the initiation of the secondary metabolism (by an entirely unknown mechanism), without the requirement of ppGpp. In fact, as reported by Allen and Noller (1), mutations in ribosomal proteins S12 and S4 may influence the higher-order structure of 16S rRNA. However, we cannot rule out the possibility that what we are observing is a novel regulation system, mediated by ribosomes, and not directly related with translation. We are now attempting to clarify, by analyzing protein and RNA formation and by studying mRNA and ppGpp levels, (i) whether certain rpsL alleles induce a stress response or not and (ii) how the combination of rpsL and rel affects these parameters.