The Gene Encoding RNase III in Streptomyces coelicolor Is Transcribed during Exponential Phase and Is Required for Antibiotic Production and for Proper Sporulation

Abstract

Phenotypic analysis of a constructed RNase III null mutant of Streptomyces coelicolor revealed that RNase III is required for both antibiotic production and proper formation of sporulation septa. Transcriptional analysis of the gene encoding RNase III indicated that it is transcribed exclusively during exponential phase as part of a tricistronic message.

Streptomyces is an enormous genus of gram-positive bacteria perhaps best known for the production of secondary metabolites with antibiotic properties. The production of antibiotics is part of the complex Streptomyces life cycle and often coincides with differentiation of vegetative hyphae into sporogenous, aerial hyphae (6). When grown in shaken liquid cultures, streptomycetes often begin to produce antibiotics during the transition from exponential to stationary phase (6). Streptomyces coelicolor A3(2) has proven to be a useful model organism for studies of antibiotic production because two of the antibiotics that it produces (actinorhodin and undecylprodiginine) are pigmented and thus easily identifiable. Among the many known S. coelicolor mutants with defects in the production of antibiotics is the absB mutant identified by Champness and coworkers (2, 9, 17). The S. coelicolor absB mutant forms aerial hyphae and spores, but it is unable to produce actinorhodin, the prodigines, and the calcium-dependent antibiotic. This mutant was found to harbor a loss-of-function mutation in a gene encoding an ortholog of Escherichia coli RNase III (rnc), a double-stranded RNA-specific endoribonuclease (5). In E. coli, this enzyme is known to process rRNA, bacteriophage RNA, and a few mRNAs, including the dicistronic transcript encoding rpsO and pnp (4, 18, 21). By analogy with E. coli, S. coelicolor RNase III is presumed to process rRNA and has been shown to cleave the rpsO-pnp transcript in vitro (10). A clear explanation for the requirement for RNase III for antibiotic production in S. coelicolor has not yet been established (1, 10).

Here, we report the construction of an S. coelicolor rnc null mutant and phenotypic characterization of this mutant. In addition, we used high-resolution S1 nuclease mapping experiments to determine when and how the gene encoding RNase III gene is transcribed.

Phenotypic analysis of an S. coelicolor rnc null mutant.

PCR-targeted mutagenesis was used for replacement of the S. coelicolor rnc gene with an apramycin resistance marker, apr (13). The requisite PCR product was amplified from the apramycin resistance gene insert of pIJ773 using primers SCO5572 KO FOR (AGGTCCTCGAGGTCTGAGCGGCTGGTGAGAGGCACTGTGATTCCGGGGATCCGTCGACC) and SCO5572 REV (TGCCGGGGCGGGCGTTCGGACCGTGCGGTGGACGGGTCATGTAGGCTGGAGCTGCTTC). The PCR product was introduced into E. coli BW25113/pIJ790 harboring cosmid St7A1 and expressing λ RED recombinase. The resultant recombinant cosmid, St7A1 Δrnc::apr, was introduced via E. coli strain ET12567/pUZ8002 into S. coelicolor M600 by conjugation. M600 is a plasmid-free derivative of the wild-type strain (15). Exconjugants lacking the rnc gene were identified by selection for apramycin resistance and kanamycin sensitivity. Gene replacement was confirmed by PCR analysis of both the recombinant cosmid and genomic DNA from the null mutant, S. coelicolor J3410 Δrnc::apr.

As expected, the phenotype of the constructed S. coelicolor rnc null mutant closely resembled that reported for the absB mutant. The null mutant did not produce actinorhodin or the prodiginines when it was grown on any of the nine different solid media tested (Fig. 1). These results are notable because the phenotypes of other mutants with pleiotropic defects in antibiotic production are dependent on the growth medium (8). Furthermore, the null mutant did not produce antibiotics when it was grown in SMM, a minimal liquid medium optimized for antibiotic production in S. coelicolor (15). After 48 h of growth in this medium, the null mutant did not contain measurable quantities of actinorhodin or the prodiginines, while wild-type S. coelicolor grown under identical conditions produced these antibiotics copiously (data not shown). In this medium, the null mutant and wild-type strain have comparable growth kinetics, but the null mutant accumulates about 20% more biomass than the parent strain by late stationary phase (Fig. 2).

FIG. 1.

The S. coelicolor rnc gene is absolutely required for antibiotic production. Wild-type S. coelicolor M600, S. coelicolor J3410 Δrnc::apr, S. coelicolor J3410/pJS72, and S. coelicolor J3410/pJS72+IRS (clockwise from the top left) were grown on minimal, complex, and chemically defined Evans media (15, 20). Ectopic expression of the S. coelicolor rnc gene with and without the 3′ inverted repeat sequence (IRS) causes overproduction of actinorhodin on SMMS, R2, and all chemically defined Evans media.

FIG. 2.

Comparison of growth kinetics of wild-type S. coelicolor and the rnc null mutant. Both strains were grown as shaken liquid cultures in SMM from an initial inoculum of ∼10⁸ CFU/ml for 36 h at 30°C. Production of the pigmented antibiotics was visible by 24 h in the wild-type cultures. The growth curves reflect the averages of three separate experiments.

In addition to the defect in antibiotic production, the constructed rnc null mutant and the absB mutant are distinguished by colonies that are about one-half the diameter of wild-type S. coelicolor colonies (2). The surfaces of the null mutant colonies were analyzed by scanning and transmission electron microscopy and by fluorescence microscopy. In electron micrographs of the null mutant colonies, it was readily apparent that there was a defect in sporulation (Fig. 3). Whereas the wild-type strain produced spores of uniform size and shape, the rnc null mutant produced both elongated and abnormally small spores. The irregularity in size indicates that RNase III is required for proper formation of sporulation septa in the aerial hyphae. It is noteworthy that both the small and elongated spores of the rnc null mutant contained DNA, as observed in fluorescent micrographs of aerial hyphae stained with propidium iodide or 4′,6-diamidino-2-phenylindole (DAPI) (data not shown). Based on spore counts followed by quantification of CFU on solid media, more than 90% of the null mutant spores were indeed viable.

FIG. 3.

The S. coelicolor rnc gene is required for proper formation of sporulation septa: scanning electron micrographs of the rnc null mutant and the wild-type strain grown for 5 days on MS medium at 30°C.

The S. coelicolor rnc gene (SCO5572) and the two upstream genes (SCO5570 and SCO5571) lie in the same orientation and are separated by only 2 and 19 bp, respectively, suggesting that the three genes might form an operon (Fig. 4A). SCO5570 encodes a hypothetical protein with a predicted metal and nucleic acid binding domain, and SCO5571 encodes ribosomal protein L32 (rpmF) (5; http://streptomyces.org.uk). For complementation, restriction fragments containing only rnc (SCO5572) or rnc together with the two upstream genes (SCO5570 and SCO5571) (Fig. 4A) were cloned into the EcoRV site of pMS81, a hygromycin-resistant vector that integrates site specifically into the chromosome at the φBT1 attP site (12). These constructs were introduced into the S. coelicolor rnc null mutant via E. coli strain ET12567/pUZ8002 by conjugation. Only provision of the rnc gene in cis with the two upstream genes restored antibiotic production (Fig. 4B), suggesting that the rnc gene is the last gene in a tricistronic message. Additional complementation experiments were designed to verify that the rnc gene alone was sufficient to restore the wild-type phenotype. The S. coelicolor rnc gene (both with and without its 3′ inverted repeat sequence) was placed under the control of the strong, constitutive ermE promoter in the hygromycin-resistant, attP integrative vector pIJ10275 (H.-J. Hong, unpublished). The gene was amplified by PCR from cosmid St7A1 using primer SCO5572 FOR (AACATATGTCAGTCCCCAAGAAGGC) and primer SCO5572 REV (AAGGATCCAAGCTTTCAGGCGGAGGCGGA) or SCO5572IRS REV (AAGGATCCAAGCTTGCCGGTGGATGAGCGA) (engineered NdeI and BamHI sites are underlined), cloned into the EcoRV site of pBluescript KS+ (Stratagene), and sequenced. NdeI and XhoI fragments from the cloning constructs were ligated into complementary sites of pIJ10275. The resulting plasmids (pJS72 and pJS72IRS) were introduced into the rnc null mutant of S. coelicolor by conjugation. In the null mutant, constitutive expression of the rnc gene not only restored antibiotic production but also caused overproduction of actinorhodin on certain solid media (Fig. 1). All complementation plasmids that restored normal antibiotic production to the rnc null mutant also restored proper formation of sporulation septa (data not shown).

FIG. 4.

Complementation of the S. coelicolor rnc null mutant. (A) Inserts of S. coelicolor J3410 (Δrnc) complementing clones and comparison to the equivalent regions of the S. coelicolor chromosome (5; http://streptomyces.org.uk). rnc is SCO5572. SCO5570 encodes a protein with a predicted metal and nucleic acid binding domain, and SCO5571 encodes ribosomal protein L32 (rpmF). The vector in all of the clones is pMS81 (12). (B) Top and bottom views of plates showing complementation of S. coelicolor J3410 (Δrnc) by pJS3 and pJS4 but not by pJS1. The cultures were grown on Difco nutrient agar for 4 days at 30°C.

Transcriptional analysis of the S. coelicolor gene encoding RNase III.

The results of the complementation experiments suggested that the rnc gene forms part of an operon (Fig. 4). To confirm this and to assess the timing of transcription, high-resolution S1 nuclease protection analysis of RNA isolated from S. coelicolor grown in liquid SMM (Fig. 2) and on R2YE solid medium (data not shown) was performed. RNA was isolated at time points in the exponential, transition, and stationary phases of growth by using previously described methods (19). In each experiment, 30 μg of RNA was hybridized with 1 fmol (10³ Cerenkov counts min⁻¹) of the probe (15, 19). The probes were amplified from the cosmid St7A1 template using the following primer pairs: SCO5570 S1-FOR (GACCAGCCTCTTCGACACCAG) and SC5570 S1-REV (GCTCGTGCGTGTCGACAC), SCO5571 S1-FOR (GCTTGTTGTAAGTGCCGCAAGC) and SCO5571 S1-REV (GCTTGTTGTAAGTGCCGCAAGC), and SCO5572 S1-FOR (TTACAACAAGCGCCAGGTCC) and SCO5572 S1-REV (GTAGGAACGGTGGGTCAGTG). The PCR products were cloned into pBluescript KS+ (Stratagene) to enable preparation of S1 nuclease probes with 75-bp 3′ nonhomologous ends as internal controls. Thus, the probes were amplified by PCR using the cognate reverse primers ³²P labeled at the 5′ end with T4 kinase and [γ-³²P]ATP and the pBluescript T3-specific primer (GCGCAATTAACCCTCACTAAAGGG). When annealed to mRNA, the probes spanning the region surrounding the translation start sites of the rnc gene and the upstream gene SCO5571 encoding ribosomal protein L32 were fully protected from the S1 nuclease; however, significant S1 nuclease digestion of the SCO5570 probe was observed, yielding a 140-nucleotide protected DNA fragment (Fig. 5A). These results indicate that the three genes are cotranscribed in the exponential phase and that the transcription start site is upstream of SCO5570. Electrophoresis of the 140-nucleotide protected DNA fragment with a sequencing ladder (generated using the same radiolabeled oligonucleotide primer that was used to generate the S1 nuclease mapping probe) indicated that the most likely transcriptional start site is 82 bp upstream of the SCO5570 start codon (Fig. 5B). The putative −35 and −10 sequences of the SCO5570 promoter are TTGGGC-N₁₈-TATCCT. This promoter is very likely to be recognized by the HrdB principal essential sigma factor (7).

FIG. 5.

Transcriptional analysis of the S. coelicolor rnc (SCO5572) gene and the two upstream genes, SCO5570 and SCO5571. (A) S1 nuclease mapping of the SCO5570, SCO5571, and SCO5572 mRNA transcripts among total RNA isolated at various time points from shaken liquid cultures of S. coelicolor M600 grown in liquid SMM at 30°C. The lanes are numbered from left to right. Lane 1 contained the φX174/HinfI ladder. Lane 2 contained the undigested probes including 85-bp nonhomologous ends derived from the T3 site of pBluescript SCO5570 (428 bp), SCO5571 (320 bp), and SCO5572 (274 bp). Lane 3 contained the digested probes mixed with a tRNA control. Lanes 4 to 9 contained the S1 nuclease-protected probe-RNA hybrids. The three probes were simultaneously hybridized to total RNA. (B) S1 nuclease mapping of the transcriptional start site of SCO5570 was conducted using a 450-bp PCR product uniquely labeled at the 5′ end. The asterisks indicate the probable transcriptional start sites. Lanes G, A, T, and C contained sequence ladders derived from the same labeled primer that was used to generate the PCR product.

Concluding remarks.

Although it does not catalyze biosynthetic reactions, RNase III is an essential enzyme for antibiotic production in S. coelicolor. The results reported here indicate that this enzyme is also required for proper formation of sporulation septa. These observations are consistent with pleiotropic effects; indeed, the E. coli rnc gene influences the level of ∼10% of all cellular proteins (11). The apparent exponential-phase transcription of the S. coelicolor rnc gene and cotranscription with the upstream gene encoding ribosomal protein L32 are consistent with a role for S. coelicolor RNase III in the processing of rRNA, as is the case for its ortholog in E. coli. Curiously, the nearly identical vegetative growth kinetics of the S. coelicolor rnc null mutant and the wild-type strain in liquid culture raise questions about the significance of rRNA processing by RNase III in S. coelicolor. It is possible that RNase III is one of many cooperative proteins that regulate antibiotic production (6, 14, 16). For instance, RNase III could catalyze a hydrolysis reaction that inactivates an mRNA encoding a repressor or activates a transcript encoding an activator of antibiotic production (1, 10). Alternatively, the phenotype of the S. coelicolor rnc null mutant could be explained by incomplete processing of rRNA that yields ribosomes lacking the requisite processivity for translation of the atypically long mRNAs of secondary metabolism genes in S. coelicolor (5; http://streptomyces.org.uk). Biochemical experiments are needed to distinguish between these two possibilities.