Antibiotic Overproduction by rpsL and rsmG Mutants of Various Actinomycetes

Abstract

Certain streptomycin resistance mutations (i.e., rpsL and rsmG) result in the overproduction of antibiotics in various actinomycetes. Moreover, rpsL rsmG double-mutant strains show a further increase in antibiotic production. rpsL but not rsmG mutations result in a marked enhancement of oligomycin production in Streptomyces avermitilis and erythromycin production in Saccharopolyspora erythraea, accompanied by increased transcription of a key developmental regulator gene, bldD, in the latter organism.

Recently, we described a practical method of increasing antibiotic production in bacteria by modulating ribosomal components (ribosomal proteins or rRNA), specifically by the introduction of mutations conferring drug resistance (5, 22, 23). This approach, called “ribosome engineering” (12, 14), has several advantages, including the ability to screen for drug resistance mutations by simple selection on drug-containing plates, even if the mutation frequency is extremely low (e.g., <10⁻¹⁰). As demonstrated with Streptomyces coelicolor A3(2) and Streptomyces albus, ΔrsmG and rpsL(K88E) mutants exhibit enhanced protein synthetic activity during the late growth phase (10, 16, 21), which is consistent with our hypothesis (4)that the capacity of a cell to synthesize proteins during the late growth phase is indicative of its ability to accelerate the onset of secondary metabolism and to produce biosynthetic enzymes. Although earlier work was conducted mainly with S. coelicolor, we have now demonstrated with various actinomycetes that rpsL and rsmG mutations are effective at enhancing the production of antibiotics and that the presence of the resistance mutations in combination (rpsL rsmG) further activates antibiotic production, as had been shown previously with S. coelicolor (10).

Bacterial strains.

The wild-type strains S. coelicolor A3(2) (actinorhodin producer), Streptomyces antibioticus strain 3720 (actinomycin producer), Streptomyces parvulus ATCC 12434 (actinomycin producer), Streptomyces avermitilis K139 (avermectin and oligomycin producer), and Saccharopolyspora erythraea NRRL2338 (erythromycin producer) were used. Spontaneous streptomycin-resistant mutants were obtained as colonies that grew within 5 to 10 days after spores were spread on GYM agar (11) containing various concentrations of streptomycin. Mutations in the rpsL or rsmG gene were determined by DNA sequencing using the primers listed in Table S1 in the supplemental material. An aveA1 deletion mutant strain of S. avermitilis, in which a region from nt 1137515 to 1139954 of the wild-type chromosome was replaced by vph of Streptomyces vinaceus, was constructed by a procedure similar to that used for the construction of aveR deletion mutants as described previously (9).

Effects of rpsL mutations on antibiotic production.

Spontaneous mutants possessing high-level (≥20-fold MIC) streptomycin resistance arose at a low frequency of around 10⁻¹⁰. The majority of resistant mutants tested had a mutation in rpsL, which encodes the ribosomal protein S12, as determined by DNA sequencing analysis (Tables 1 and 2). Actinomycin-overproducing strains were found at a high incidence of 5% to 20% among the rpsL mutants derived from S. antibioticus and S. parvulus, showing a three- to fivefold increase in actinomycin production (Table 1). Similarly, the introduction of rpsL mutations effectively increased the antibiotic production of S. avermitilis (Fig. 1) and S. erythraea (Fig. 2A). The introduction of certain rpsL mutations (e.g., K43M and K88E) markedly increased the production of oligomycin (20- to 40-fold) in S. avermitilis. Interestingly, rpsL mutations always resulted in a decrease in avermectin production. This was especially pronounced in rpsL mutations that effectively increased oligomycin production. This may be explained by competition between oligomycin biosynthesis and avermectin biosynthesis for common precursors (i.e., malonyl coenzyme A and methylmalonyl coenzyme A) that are the building blocks of their polyketide moieties. In fact, a mutation of the aveA1 gene (the wild-type form of which codes for avermectin-aglycone polyketide synthase) that leads to complete abolition of avermectin production also caused a marked increase in oligomycin production, reaching 503 μg/ml. As expected, the introduction of the rpsL(K43M) mutations into the blocked mutant was also effective at increasing oligomycin production (reaching 1,064 μg/ml), demonstrating the direct effect of the rpsL mutations on the activation of oligomycin production.

TABLE 1.

Characterization of streptomycin resistance mutations of S. antibioticus, S. parvulus, and S. coelicolor

Strain	Streptomycin concn (μg/ml) used for selection	Concn of indicated antibiotic (μg/ml)a	Mutation inb:		Amino acid substitution	Resistance to streptomycin (μg/ml)c
			rpsL	rsmG
		Actinomycin
S. antibioticus 3720	—d	5.5 ± 0.6				0.5
YT1	20	27.7 ± 5.4	262A→G		Lys88→Glu	30
YT2	20	19.9 ± 1.2	263A→G		Lys88→Arg	30
YT3	20	22.1 ± 5.2	262A→G		Lys88→Glu	30
YT4	20	15.1 ± 6.3	263A→G		Lys88→Arg	30
YT5	1	31.2 ± 4.4		251C→G	Pro84→Arg	1
YT11	20	46.2 ± 6.7	262A→G (YT5 + rpsL)		Lys88→Glu	>200
YT12	20	47.1 ± 4.4	263A→G (YT5 + rpsL)		Lys88→Arg	>200
S. parvulus ATCC 12434	—	6.2 ± 2.1				2
YT20	50	19.2 ± 2.3	128A→G		Lys43→Arg	>200
YT21	50	32.8 ± 8.6	263A→G		Lys88→Arg	100
YT22	50	13.8 ± 4.4	128A→G		Lys43→Arg	>200
YT23	3	14.5 ± 5.0		233C→T	Ala78→Val	10
YT24	3	21.1 ± 9.3		239T→C	Leu80→Pro	10
		Actinorhodin
S. coelicolor 1147	—	1.8 ± 0.5				1
KO-178	100	70.2 ± 5.8	262A→G		Lys88→Glu	100
KO-179	3	46.4 ± 0.5		21C→CC	Frameshift	5
KO-944	100	133.8 ± 7.0	257G→C (KO-179 + rpsL)		Arg86→Pro	200
KO-945	100	131.8 ± 10.5	262A→G (KO-179 + rpsL)		Lys88→Glu	500

^aActinomycin was determined as described previously (13) after 4 days (for S. antibioticus) or 5 days (for S. parvulus) of cultivation at 30°C in SYM medium (13) for S. antibioticus or 2× SYM medium for S. parvulus. Actinorhodin was determined by the method of Kieser et al. (8) after 8 days of cultivation at 30°C in GYM medium. All measurements were performed in triplicate.

^bNumbered from the start codon of the open reading frame (in S. coelicolor) or in accordance with the numbering system for S. coelicolor (in S. antibioticus and S. parvulus).

^cDetermined after 3 days of incubation on GYM medium.

^d—, wild-type strain.

TABLE 2.

Locations of rpsL and rsmG gene mutations and resulting amino acid changes in S. avermitilis and S. erythraea

Strain	Streptomycin concn (μg/ml) used for selection	Mutation ina:		Amino acid substitution	Frequency of mutants with same mutation	Resistance to streptomycin (μg/ml)b
		rpsL	rsmG
S. avermitilis K139	—c					0.2
KO-770	10	128A→C		Lys43→Thr	3/34	5,000
KO-772	10	129G→C		Lys43→Asn	1/34	5,000
KO-774	10	129G→T		Lys43→Asn	9/34	5,000
KO-782	10	128A→G		Lys43→Arg	1/34	1,000
KO-783	10	128A→T		Lys43→Met	1/34	5,000
KO-776	10	262A→G		Lys88→Glu	11/34	20
KO-780	10	263A→T		Lys88→Met	4/34	20
KO-778	10	263A→G		Lys88→Arg	4/34	10
KO-798	0.5		429C→CC	Frameshift	1/25	0.6
KO-799	0.5		316A→G	Glu106→Lys	3/25	0.6
KO-800	0.5		344T→G	Leu115→Arg	1/25	0.6
KO-801	0.5		515T→TAGAT	Frameshift	1/25	0.6
KO-802	0.5		553A→T	Lys185→stop codon	1/25	0.6
KO-964			ΔrsmGd			0.6
S. erythraea NRRL2338	—					8
KO-950	50	122C→T		Thr41→Ile	2/72	100
KO-951	50	124C→G		Pro42→Ala	1/72	300
KO-952	50	128A→G		Lys43→Arg	8/72	3,000
KO-953	50	128A→T		Lys43→Met	1/72	5,000
KO-954	50	129G→T		Lys43→Asn	1/72	>5,000
KO-955	50	132G→C		Lys44→Asn	2/72	200
KO-956	50	256C→T		Arg86→Cys	2/72	300
KO-957	50	263A→G		Lys88→Arg	18/72	2,000
KO-958	50	272C→A		Pro91→His	2/72	500
KO-959	50	274G→C		Gly92→Arg	2/72	800
KO-960	20		8G→A	Gly3→Asp	1/24	50
KO-961	20		Δ38T-48C	Frameshift	1/24	50
KO-962	20		56A→T	Glu19→Val	1/24	50
KO-963	20		200T→C	Ala67→Val	1/24	50

^aNumbered in accordance with the numbering system for S. coelicolor.

^bDetermined after 3 days (for S. avermitilis) or 5 days (for S. erythraea) of incubation on GYM medium.

^c—, wild-type strain.

^dThe rsmG gene of strain K139 was disrupted by genetic engineering as described in the legend to Fig. 1.

FIG. 1.

Effects of introducing rpsL or rsmG mutations on production of oligomycin and avermectin in S. avermitilis. The disruption of rsmG in S. avermitilis was performed as follows. A 3.4-kbp fragment spanning rsmG was amplified by PCR using primers rsmG-F3 (5′-CAGAATTCGTGAAGACGCCGTTCATGCA-3′) and rsmG-R3 (5′-CAGAATTCCACCATGCTCTGTTCGTTC-3′), each containing an EcoRI site (underlined). The PCR fragment was inserted into the EcoRI site of pGEM-9Zf(−) (Promega), yielding pGEMrsmG. An ermE gene fragment was amplified from pIJ4026 using primers ermE-F and ermE-R (see Table S1 in the supplemental material) and ligated with pGEMrsmG that had been treated with NruI and BamHI using the In-fusion dry-down PCR cloning kit (Clontech). Then, the ΔrsmG::ermE deletion construct was shuttled into pGM160::oriT to yield pGMΔrsmG::ermE. The recombinant plasmid was transformed into E. coli ET12567/pUZ8002 and introduced into S. avermitilis wild-type strain K139. Exoconjugants were selected using thiostrepton and lincomycin. S. avermitilis ΔrsmG mutants were selected as described previously (10) and identified as colonies with thiostrepton-sensitive and lincomycin-resistant phenotypes. Correct disruption of rsmG was confirmed by PCR using the primers rsmG-F1 and rsmG-R1 (see Table S1 in the supplemental material). The culture conditions of S. avermitilis for production of oligomycin and avermectin (shown for each chemical structure at the right) and assay conditions for these antibiotics were described previously (6, 7). Error bars indicate the standard deviations of the means of triplicate or more samples.

FIG. 2.

Effects of introducing rpsL or rsmG mutations on erythromycin production in S. erythraea NRRL2338. (A) Erythromycin production by various mutant strains. Strains were grown in CFM1 medium (17) at 30°C for 6 days. Erythromycin production was determined by bioassay (agar diffusion method) using S. aureus 209P as a test organism. Error bars represent standard deviations. (B) Transcriptional analysis of bldD by real-time quantitative PCR (qPCR). Total RNA preparation and real-time qPCR were performed as described by Wang et al. (22), except that random hexamers were used in the reverse transcription reaction. Each transcriptional assay was normalized to the corresponding transcriptional level of sigA, a gene encoding the principal sigma factor. Primers used for real-time qPCR are listed in Table S1 in the supplemental material. The error bars indicate the standard deviations of the means of triplicate or more samples.

In contrast to Streptomyces spp., streptomycin-resistant mutants of S. erythraea often displayed novel mutations (e.g., T41I and P42A) in the rpsL gene that have not been previously reported (Table 2). Moreover, the K88E mutation that has been frequently found in Streptomyces spp. was not detected. Among the identified rpsL mutations, the K43N mutation was particularly effective at enhancing erythromycin production (twofold increase) (Fig. 2A). Although the erythromycin biosynthetic cluster lacks a pathway-specific regulatory gene, recent work (3) has demonstrated that a key developmental regulator encoded by the bldD gene positively regulates the synthesis of erythromycin. As expected, the rpsL(K43N) mutant KO-954 exhibited a higher level of bldD expression at the late growth phase (60 h) than that of the wild-type strain (Fig. 2B), thus underlying the enhanced production of erythromycin in this mutant.

Effects of rsmG mutations.

As shown in Tables 1 and 2, nearly half of the low-level (≥2-fold MIC) streptomycin-resistant mutants possessed an rsmG mutation. Mutations in rsmG were effective at enhancing actinomycin production (in S. antibioticus and S. parvulus) and actinorhodin production (in S. coelicolor) (Table 1), whereas rsmG mutations were only moderately effective at enhancing oligomycin production in S. avermitilis, as determined using an rsmG-disrupted mutant, KO-964 (Fig. 1). In contrast, rsmG mutations did not result in enhancement of erythromycin production in S. erythraea (Fig. 2A), indicating the species dependence of the rsmG mutation effect. In S. coelicolor, enhanced expression of the metK gene encoding S-adenosylmethionine (SAM) synthetase corresponds to the enhanced production of actinorhodin caused by rsmG mutations (10, 15). Accordingly, the rsmG mutant of S. antibioticus (YT5) showed three- to fivefold greater activity of SAM synthetase (see Fig. S1 in the supplemental material), while the S. erythraea rsmG mutant KO-961 displayed no increase in SAM synthetase activity (and no increase in the transcription level of metK) (data not shown), indicating the lack of an equivalent effect of rsmG mutations in S. erythraea.

Antibiotic overproduction in rpsL rsmG double mutants.

Combining the rpsL and rsmG mutations resulted in a further increase in antibiotic production, as determined by constructing rpsL rsmG double mutants by generating high-level streptomycin-resistant mutants from rsmG mutants. The resulting high-level streptomycin-resistant mutants were found to possess a point mutation in the rpsL gene, causing a novel amino acid alteration at position Lys43, Lys45, Arg86, or Lys88 (data not shown). In S. coelicolor, the rpsL(R86P) mutation, in addition to the rpsL(K88E) mutation (10), was found to be effective at further increasing actinorhodin production in combination with rsmG mutations (Table 1). In addition, rpsL rsmG double-mutant strains of S. antibioticus displayed a further increase in actinomycin production (Table 1), demonstrating the efficacy of rpsL rsmG double mutations on antibiotic overproduction.

Concluding remarks.

Secondary metabolism is often controlled by pathway-specific regulatory genes (e.g., actII-ORF4 for actinorhodin production and strR for streptomycin production) (1, 18). It is possible that the expression of pathway-specific regulatory genes is governed by higher-order regulatory proteins and that expression of such higher-order regulatory proteins may be significantly affected under conditions associated with enhanced protein synthesis during the stationary phase in mutants. The enhanced expression of bldD in the S. erythraea rpsL mutant at the transcriptional level (Fig. 2B) may be explained in the same way, although an S. erythraea rif mutant with increased productivity of erythromycin does not show an increased transcription of bldD (2). It should also be pointed out that rpsL mutations had opposite effects on the production of oligomycin and avermectin, eventually leading to “preferential” biosynthesis of oligomycin (Fig. 1). The rpsL mutations may therefore be effective for switching to production of the desired antibiotic. The present method, together with the methods reported recently (4a, 19, 20, 22), may be useful as a practical approach for strain improvement.