ABSTRACT
Pristinamycin I (PI), produced by Streptomyces pristinaespiralis, is a streptogramin type B antibiotic, which contains two proteinogenic and five aproteinogenic amino acid precursors. PI is coproduced with pristinamycin II (PII), a member of streptogramin type A antibiotics. The PI biosynthetic gene cluster has been cloned and characterized. However, thus far little is understood about the regulation of PI biosynthesis. In this study, a TetR family regulator (encoded by SSDG_03033) was identified as playing a positive role in PI biosynthesis. Its homologue, PaaR, from Corynebacterium glutamicum serves as a transcriptional repressor of the paa genes involved in phenylacetic acid (PAA) catabolism. Herein, we also designated the identified regulator as PaaR. Deletion of paaR led to an approximately 70% decrease in PI production but had little effect on PII biosynthesis. Identical to the function of its homologue from C. glutamicum, PaaR is also involved in the suppression of paa expression. Given that phenylacetyl coenzyme A (PA-CoA) is the common intermediate of the PAA catabolic pathway and the biosynthetic pathway of l-phenylglycine (l-Phg), the last amino acid precursor for PI biosynthesis, we proposed that derepression of the transcription of paa genes in a ΔpaaR mutant possibly diverts more PA-CoA to the PAA catabolic pathway, thereby with less PA-CoA metabolic flux toward l-Phg formation, thus resulting in lower PI titers. This hypothesis was verified by the observations that PI production of a ΔpaaR mutant was restored by l-Phg supplementation as well as by deletion of the paaABCDE operon in the ΔpaaR mutant. Altogether, this study provides new insights into the regulation of PI biosynthesis by S. pristinaespiralis.
IMPORTANCE A better understanding of the regulation mechanisms for antibiotic biosynthesis will provide valuable clues for Streptomyces strain improvement. Herein, a TetR family regulator PaaR, which serves as the repressor of the transcription of paa genes involved in phenylacetic acid (PAA) catabolism, was identified as playing a positive role in the regulation of pristinamycin I (PI) by affecting the supply of one of seven amino acid precursors, l-phenylglycine, in Streptomyces pristinaespiralis. To our knowledge, this is the first report describing the interplay between PAA catabolism and antibiotic biosynthesis in Streptomyces strains. Considering that the PAA catabolic pathway and its regulation by PaaR are widespread in antibiotic-producing actinomycetes, it could be suggested that PaaR-dependent regulation of antibiotic biosynthesis might commonly exist.
INTRODUCTION
Pristinamycin I (PI), produced by Streptomyces pristinaespiralis, is a branched cyclohexadepsipeptide antibiotic and belongs to the B group of streptogramins. PI is coproduced with pristinamycin II (PII), a polyunsaturated cyclopeptidic macrolactone antibiotic, which is a member of the A group of streptogramin antibiotics (1, 2). Two substances are produced in a 30:70 ratio and show a potent synergistic effect with an approximately 100-fold-higher bactericidal activity than that from treatment with a single component alone (3). The combined application of PI and PII semisynthetic derivatives, quinupristin-dalfopristin, is very active against many Gram-positive bacteria, including multidrug-resistant pathogens, such as methicillin-resistant Staphylococcus aureus (MRSA), drug-resistant Streptococcus pneumoniae, and vancomycin-resistant Enterococcus faecium (4).
PI biosynthesis is catalyzed by a nonribosomal peptide synthetase (NRPS) complex composed of SnaA, SnaC, and SnaDE, which are responsible for the successive condensation of two proteinogenic amino acids, l-threonine and l-proline, and five aproteinogenic amino acids, including 3-hydroxypicolinic acid, l-aminobutyric acid, 4-oxo-l-pipecolic acid, l-phenylglycine (l-Phg), and 4-N,N-dimethylaminol-phenylalanine (DMAPA) or N-methyl-4-(methylamino)-l-phenylalanine (MMAPA) (5). In the PI biosynthetic gene cluster, 12 genes are involved in PI precursor supply, including hpaA, pipA and snbF and gene clusters papACBM and pglABCDE, which are required for the formation of 3-hydroxypicolinic acid, 4-oxo-l-pipecolic, DMAPA (or MMAPA), and l-Phg, respectively (5, 6). Although a possible biosynthetic pathway has been proposed, little is understood about the regulation of PI biosynthesis in S. pristinaespiralis. To our knowledge, until now, only seven cluster-situated regulatory genes (from papR1 to papR6 and spbR) and a serine/threonine protein kinase gene, spy1, were identified as being involved in the regulation of PI biosynthesis (2, 5, 7–9). For the genus Streptomyces, manipulating the regulatory network that controls antibiotic biosynthesis has been proven to be an efficient method for improving the production levels of desired antibiotics (10). Therefore, a better understanding of the regulation mechanisms for PI biosynthesis will provide valuable clues for strain improvement to increase PI titer by metabolic engineering approaches.
In our previous work, we generated a dozen S. pristinaespiralis mutants with the individual deletion of a number of TetR family regulatory genes and have checked their effects on pristinamycin production (unpublished data). A mutant with significantly reduced PI production was that with the deletion of SSDG_03033. The regulator encoded by SSDG_03033 is highly conserved in Actinobacteria and shows a high amino acid sequence identity with PaaR from Corynebacterium glutamicum, which functions as a repressor of the phenylacetic acid (PAA) degradation pathway (11). Thus, we also named the SSDG_03033-encoding regulator PaaR. Herein, the mechanism underlying its function in PI production was revealed. We demonstrated that PaaR plays a positive role in PI biosynthesis by repressing the transcription of paa genes involved in PAA catabolism and thereby possibly diverting phenylacetyl coenzyme A (PA-CoA) to the biosynthetic pathway of l-Phg, the last building block for PI biosynthesis. PA-CoA is the common intermediate for both the PAA degradation pathway and the l-Phg biosynthetic pathway, as shown in Fig. 1 (6, 11).
Schematic presentation of the proposed phenylacetic acid (PAA) degradation pathway and l-Phg biosynthetic pathway. The PAA degradation pathway and l-Phg biosynthetic pathway are indicated by blue and red arrows, respectively. PglA, hydroxyacyl-dehydrogenase; PglB, pyruvate dehydrogenase E1 component α-subunit; PglC, pyruvate dehydrogenase E1 component β-subunit; PglD, thioesterase type II; PglE, phenylglycine aminotransferase; SnbE, pristinamycin I (PI) synthase 3 and 4; PaaABCDE, ring 1,2-phenylacetyl-CoA epoxidase subunits; PaaF, 2,3-dehydroadipyl-CoA hydratase; PaaG, ring 1,2-epoxyphenylacetyl-CoA isomerase, oxepin-CoA forming; PaaH, 3-hydroxyadipyl-CoA dehydrogenase; PaaJ, 3-oxoadipyl-CoA/3-oxo-5,6-dehydrosuberyl-CoA thiolase; PaaK, phenylacetyl-CoA ligase; PaaZ, fused oxepin-CoA hydrolase–3-oxo-5,6-dehydrosuberyl-CoA semialdehyde dehydrogenase.
MATERIALS AND METHODS
Bacterial strains, growth conditions, plasmids.The bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli strains, including DH5α, BL21(DE3), and ET12567/pUZ8002, were cultivated in Luria-Bertani (LB) medium at 37°C. DH5α was used as the host for routine molecular cloning. ET12567/pUZ8002 was used as the donor strain in the intergeneric conjugation. BL21(DE3) was employed for protein overexpression. Antibiotics, including ampicillin (100 μg/ml), apramycin (50 μg/ml), kanamycin (50 μg/ml), and/or chloramphenicol (25 μg/ml), were added to the medium when appropriate.
Strains and plasmids used in this study
S. pristinaespiralis HCCB10218 (CGMCC 5486), which was isolated after physical and chemical treatments of ATCC 25486, was used as the original strain. S. pristinaespiralis strains were grown at 30°C in RP liquid medium (peptone, 5; yeast extract, 5; valine, 0.5; NaCl, 2; KH2PO4, 0.5; MgSO4·7H2O, 1[in grams per liter]; pH 6.4) for genomic DNA isolation (8). RP agar medium was used for the preparation of spore suspensions and conjugal transfer between E. coli and S. pristinaespiralis (12). For the S. pristinaespiralis fermentation, seed medium and fermentation medium were prepared as described previously (5). When necessary, apramycin (50 μg/ml), kanamycin (50 μg/ml), and/or thiostrepton (50 μg/ml) was added.
Construction of the ΔpaaR and ΔpaaR ΔpaaABCDE deletion mutants.A ΔpaaR mutant with an in-frame deletion of the DNA sequence encoding amino acids ranging from position 38 to position 195 of the PaaR regulator was constructed on the basis of the parental strain S. pristinaespiralis HCCB10218 by traditional homologous recombination (12). The upstream and downstream regions (1,243 and 1,249 bp, respectively) of the target DNA sequence were amplified from the genomic DNA of HCCB10218 by using primer pairs paaRup-fw/rv and paaRdown-fw/rv (see Table S1 in the supplemental material). The two PCR products were digested with HindIII/XbaI and XbaI/EcoRV, respectively, and cloned into the replication temperature-sensitive plasmid pKC1139 between HindIII and EcoRV to yield pKC-paaR. The construct was introduced into ET12567/pUZ8002 and subsequently transferred to the parental strain HCCB10218 by conjugal transfer. To obtain the single-crossover strains, apramycin-resistant strains were passaged three times on RP agar plates containing apramycin at 37°C. The single crossovers were checked by PCR using the primer pair pKC1139-F/paaRSC-rv (see Table S1). The above obtained strains were inoculated into apramycin-free RP liquid medium and passaged for three rounds. Subsequently, the cultures were serially diluted and plated on apramycin-free RP agar medium. To identify the double-crossover strains, the single colonies were picked and grown on apramycin-free and -supplemented RP agar plates by replica plating. The ΔpaaR mutant strains were checked by PCR using the primer pair paaRSJ-fw/rv (see Table S1), followed by DNA sequencing.
A ΔpaaR ΔpaaABCDE double mutant with the deletion of the entire paaABCDE gene cluster based on the ΔpaaR mutant was constructed by an I-SceI-mediated homologous recombination (13). The upstream and downstream homologous arms (1,757 and 1,937 bp, respectively) were amplified with primer pairs paaA-Eup-fw/rv and paaA-Edown-fw/rv (see Table S1 in the supplemental material). The PCR products were treated with SwaI/XbaI and XbaI/PacI, respectively, and ligated to plasmid pKC1139-SPI between SwaI and PacI to yield pKC-paaABCDE (Table 1). The construct was introduced into the ΔpaaR mutant by conjugal transfer. Apramycin-resistant strains were passaged three times on RP agar supplemented with apramycin. The single-crossover events were confirmed by PCR using the primer pair pKC1139-F and paaA-ESC-rv (see Table S1). After isolation of the spores of the single-crossover strains, the plasmid pALSceI (Table 1), which contains the codon-optimized I-SceI-encoding gene under the control of an inducible promoter tipAp, was introduced by conjugal transfer and I-SceI expression was induced by thiostrepton. Thiostrepton-resistant strains were checked for double crossovers by replica plating. The mutant strains were verified by PCR using the primers paaA-ESJP-fw/rv and paaA-E-SJN-fw/rv (see Table S1). The resulting strains were passaged three times on RP agar medium without antibiotics to remove plasmid pALSceI, resulting in the ΔpaaR ΔpaaABCDE mutant.
Complementation assay.The paaR opening reading frame (ORF; 624 bp) was amplified from the genomic DNA of HCCB10218 using the primer pair paaRex-fw/rv (see Table S1 in the supplemental material). After treatment with NdeI and EcoRI, the resultant PCR product was ligated to plasmid pIB139, generating pIB-paaR, in which paaR expression was under the control of a strong constitutive promoter ermE*p. The obtained plasmid was transferred to the ΔpaaR mutant by conjugal transfer, resulting in the ΔpaaR/pIB-paaR complemented strain. Two other strains, 10218/pIB139 and the ΔpaaR/pIB139 mutant, were constructed as controls by introducing the empty plasmid pIB139 into HCCB10218 and the ΔpaaR mutant, respectively.
S. pristinaespiralis fermentation and analysis of pristinamycin production.S. pristinaespiralis fermentation and analysis of pristinamycin production were carried out according to the previously described methods with some modifications (5, 8). Briefly, the S. pristinaespiralis strains were grown on RP agar medium at 30°C for 4 to 5 days. Subsequently, the agar cultures were inoculated into 25 ml of seed medium in 250-ml flasks at 27°C and 240 rpm. After incubation for 40 to 44 h, 2 ml of the seed cultures was transferred into 25 ml of fermentation medium in 250-ml flasks. Fermentation samples (0.5 ml each) were collected at 48 and 72 h, respectively, and extracted with the same volume of acetone for 60 min. The mixtures were centrifuged at 12,000 rpm for 5 min, and pristinamycin production in the supernatants was analyzed by bioassay and high-performance liquid chromatography (HPLC). In the bioassay, Bacillus subtilis ATCC 6633 was used as the indicator strain. HPLC analysis of pristinamycin production was performed as described previously (8). The standard curves for pristinamycin quantitative analysis were made using purified PIa and PIIa (98%) obtained from Laiyi Co. Ltd. (Shanghai, China).
RNA isolation and qPCR analysis.Fermentation samples of S. pristinaespiralis were collected at 48 h. After being frozen in liquid nitrogen, the cells were ground into powder. RNA samples were prepared using an ultrapure RNA kit (Cwbio, Shanghai, China). RNase-free DNase I (TaKaRa, Dalian, China) was utilized to remove the residual genomic DNA. The integrity and quality of RNA samples were analyzed by 1.5% agarose gel electrophoresis, and the quantity was measured with a NanoDrop 2000 spectrophotometer (Thermo Scientific, USA). RNA reverse transcription was performed using Moloney murine leukemia virus (M-MLV) reverse transcriptase (Invitrogen, USA). Quantitative real-time reverse transcription (RT)-PCR (qPCR) analysis was performed as described previously (8). The primer pairs used for qPCR are listed in Table S2 in the supplemental material. The reactions were performed in a MyiQ2 PCR machine (Bio-Rad, USA) with the following parameters: DNA denaturation at 95°C for 2 min and 40 cycles of 95°C for 20 s, 60°C for 20 s, and 72°C for 20 s. PCRs were conducted in triplicate for each gene. The relative expression levels of the tested genes were normalized to hrdB (SSDG_06142, encoding a housekeeping sigma factor) and were determined using the 2−ΔΔCT method (14). The relative value for the expression of each gene in the parental strain HCCB10218 was assigned as 1. qPCR analysis was performed with three independent RNA samples (biological replicates), and error bars represent the standard deviations (SD).
Determination of the TSSs.5′-RACE (rapid isolation of cDNA end) experiments were performed using the 5′-Full RACE kit with TAP (TaKaRa, Dalian, China) according to the instructions provided by the manufacturer. Apart from the primers packaged in the kit, the gene-specific primers are listed in Table S2 in the supplemental material. After two rounds of PCR amplification (including a nested PCR), the 5′-RACE products were cloned into the pMD-18T vector (TaKaRa, Dalian, China). Ten plasmids isolated from the transformants were sequenced, and the specific transcriptional start sites (TSSs) were determined if more than seven of the sequenced cDNAs had identical 5′ ends.
Overexpression and purification of the recombinant PaaR protein.The entire paaR gene sequence was amplified from HCCB10218 genomic DNA using the primers paaRex-fw/rv (see Table S1 in the supplemental material). After treatment with NdeI and EcoRI, the PCR product was ligated to the expression vector pET28a, resulting in pET-paaR. The correct recombinant plasmids were verified by DNA sequencing. The obtained plasmid pET-paaR was transformed into E. coli BL21(DE3) competent cells. The overexpression of His6-PaaR was induced by adding IPTG (isopropyl-β-d-thiogalactopyranoside) to a concentration of 500 μM, and the cultures were grown at 16°C overnight. Cells were collected by centrifugation at 5,000 rpm for 10 min and resuspended in the binding buffer (50 mM Tris-HCl, pH 8.0, 500 mM NaCl, 10 mM imidazole, 10% glycerol). The cell suspension was disrupted using a French press (Constant Systems Limited, United Kingdom), and cell debris was removed by centrifugation. His6-PaaR was purified using a Ni-nitrilotriacetic acid (Ni-NTA)–agarose column (GE Health Care, Sweden) and eluted with increasing concentrations of imidazole from 10 mM to 500 mM. The concentration of His6-PaaR was determined using a Bradford kit (Sangon, China), and His6-PaaR purity was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).
EMSA.Electrophoretic mobility shift assays (EMSAs) between purified His6-PaaR and the tested probes (see Table S3 in the supplemental material) were carried out using a previously described method (15). Briefly, DNA fragments containing the respective promoter regions were amplified from HCCB10218 genomic DNA using the primer pairs listed in Table S1 in the supplemental material. Cy5-labeled oligonucleotide (5′-AGCCAGTGGCGATAAG-3′) was used in the second-round PCR to make the labeled probes. The labeled probes (10 ng) were incubated with different amounts of purified His6-PaaR at 25°C for 20 min in the binding buffer as described before (8). The intergenic region between SSDG_03653 and SSDG_05964 that contains no putative PaaR-binding motif was tested as a negative control. The specificity of the DNA-protein interactions was verified by adding 2 μg of unlabeled probes or salmon sperm DNA. The unlabeled DNAs were mixed with His6-PaaR at 25°C for 20 min before the labeled probes were added. Finally, the reaction mixtures were separated on 1.5% Tris-acetate-EDTA (TAE)–agarose gels in 0.5× TAE buffer. After 1 h of electrophoresis at 4°C, the gels were scanned using a FLA-9000 phosphorimager (Fujifilm, Japan).
RESULTS
Bioinformatical analysis of PaaR, a TetR family regulator encoded by SSDG_03033.The SSDG_03033 gene encodes a TetR family regulator, and its homologues are widespread in Actinobacteria. BLAST analysis revealed that this regulator exhibits high amino acid sequence identities with its homologues from other Streptomyces strains (78 to 89%; the maximum number of aligned sequences to display is set at 100) (data not shown). We show the alignment of the deduced amino acid sequences of the SSDG_03033-encoded regulator and its seven homologues from the Streptomyces whole genomes available online (http://streptomyces.org.uk/) (see Fig. S1 in the supplemental material). These eight regulators harbor nearly identical TetR N domains, which are involved in DNA binding. The functions of these homologues are yet uncharacterized. In actinomycetes, thus far, only PaaR from C. glutamicum, a homologue of the SSDG_03033-encoded regulator, has been functionally identified, which is involved in suppression of the transcription of the paa genes encoding the phenylacetic acid (PAA) catabolism pathway (11). Herein, we also name the SSDG_03033-encoded regulator PaaR.
Similar to that in C. glutamicum, the paaR gene in S. pristinaespiralis is clustered with the paa genes (Fig. 2). However, the organization of the paa gene cluster in S. pristinaespiralis is apparently different from that in C. glutamicum (11), although a coherent paaABCDE operon was found in the genome of S. pristinaespiralis. We also found that in comparison with C. glutamicum, all paa genes are clustered together in S. pristinaespiralis, some paa genes, such as paaI and paaK, are scattered elsewhere, and no paaFGJYT homologous genes were found. Interestingly, paaR in S. pristinaespiralis is located immediately adjacent to bkdR and the bkdABC operon, which encode a leucine response regulator and a branched-chain amino acid dehydrogenase complex, respectively (16). Further bioinformatical analysis revealed that the genetic organization of paa genes may be relatively conserved in different Streptomyces strains (see Fig. S2 in the supplemental material). It is worth noting that a highly conserved bkdABC-bkdR-paaR gene cluster was observed in the genomes of all selected Streptomyces strains, suggesting that PaaR is likely to participate in the regulation of branched-chain amino acid catabolism in Streptomyces.
Genetic organization of the paa gene cluster in the genomes of S. pristinaespiralis and C. glutamicum. Homologous paa genes in these two strains are indicated in the same color. The bkdR and bkdABC genes are represented by green and yellow, respectively. The genes that have no relationship with the PAA degradation pathway are represented by gray. The probes containing the corresponding intergenic regions bound by purified PaaR protein in the EMSA analysis are indicated by vertical red arrows, and the probe with no interaction with PaaR is indicated by a vertical gray arrow. The PaaR-binding sites are represented by red dots. bkdR encodes a leucine-response regulator; bkdABC encode a branched-chain amino acid dehydrogenase complex; paaT, paaI, and paaY encode a putative transporter, a thioesterase, and an acetyltransferase, respectively. The functions of other paa genes are as described in the legend for Fig. 1.
Deletion of paaR affects only PI biosynthesis.We constructed a ΔpaaR mutant with partial deletion of the paaR gene sequence based on the parental strain HCCB10218, and pristinamycin production was assayed by both a bioassay and HPLC analysis. The fermentation samples were collected at 48 and 72 h. Little difference in bacterial growth was observed between the parental strain and the ΔpaaR mutant (data not shown). The bioassay (using the supernatants extracted from fermentation samples collected at 48 h and B. subtilis as the indicator strain) showed that the inhibition zone produced by the ΔpaaR mutant is much smaller than that of the parental strain (Fig. 3A). Further quantitative analysis of pristinamycin production by HPLC revealed that deletion of paaR resulted in a 3.3-fold reduction in PIa (the major component of PI) production. However, little effect on PIIa (the major component of PII) production was detected upon paaR inactivation (Fig. 3B). Introduction of the wild-type paaR gene (under the control of a strong constitutive promoter ermE*p on an integrative plasmid pIB139) into the ΔpaaR mutant could easily restore PIa production and the inhibition zone to the levels of the parental strain HCCB10218 (with the control vector pIB139) (Fig. 3A and B). These results clearly demonstrate that PaaR is involved in activation of PI biosynthesis in S. pristinaespiralis.
Effects of paaR deletion on PI production. (A) Bioassay of pristinamycin production using B. subtilis as the indicator strain. Fermentation samples were collected at 48 h. (B) HPLC analysis of pristinamycin production. PIa and PIIa are the main components of PI and PII, respectively. Fermentation samples were collected at two time points (48 and 72 h). Error bars represent the standard deviations from three biological replicates. Five strains, including the parental strain HCCB10218, the ΔpaaR mutant, HCCB10218 containing the control vector (10218/pIB139), the ΔpaaR mutant containing the control vector (ΔpaaR/pIB139), and the complemented strain (ΔpaaR/pIB-paaR), were tested.
PaaR functions positively in PI production by affecting the transcription of the paa genes involved in PAA degradation.Given that the PaaR homologue in C. glutamicum functions as a repressor of the transcription of the paa genes and that, moreover, the PaaR-binding motif was detected in the promoter regions of the paa genes in three Streptomyces strains, including Streptomyces coelicolor, Streptomyces avermitilis, and Streptomyces griseus (11), we speculated that PaaR in S. pristinaespiralis may be also implicated in the suppression of paa gene transcription. As shown in Fig. 1, phenylacetyl-CoA (PA-CoA) is the common intermediate of the l-Phg biosynthetic pathway and PAA catabolic pathway (Fig. 1). If PaaR exerts a repressor role in paa expression as proposed above, its inactivation would enhance the transcription of these paa genes and thereby result in more consumption of PA-CoA by the PAA degradation pathway and accordingly reduce the formation of l-Phg as well as PI.
To explore this possibility, as the first step, we compared the transcription of the paa genes in the ΔpaaR mutant with that in HCCB10218 using qPCR analysis. Five paa genes, including paaH, paaZ, paaA, paaK, and paaI, were selected and tested. Four genes, including two PI biosynthetic genes (hpaA and pglC), one PII biosynthetic gene, snaN, and paaR itself, were tested as negative controls. In addition, as paaR is clustered with bkdR and bkdABC, the transcription of bkdR and bkdA was also analyzed to assess whether PaaR is involved in the regulation of these bkd genes. RNA samples were isolated from fermentation samples collected at 48 h. Transcriptional analysis revealed that the mRNA abundance of four paa genes (paaH, paaZ, paaA, and paaK) was enhanced at least 10-fold after paaR deletion. Deletion of paaR also resulted in enhanced expression of bkdA and bkdR by >2- and 4-fold, respectively. However, little difference was observed in the transcription of paaI and four PI/PII biosynthetic genes in the ΔpaaR mutant compared with that in the parental strain (Fig. 4). Moreover, as expected, no paaR transcription was detected in the ΔpaaR mutant. These data clearly demonstrated that PaaR repressed the transcription of the majority of paa genes as well as bkdA and bkdR in S. pristinaespiralis and suggested that the role of PaaR in PI biosynthesis is independent of the PI biosynthetic genes. Since paaH overlaps with paaR by 33 bp (Fig. 2), it is likely that PaaR regulates its own transcription from the promoter upstream of paaH. In addition, considering the functions of bkdABC and bkdR in S. coelicolor as described previously (16), we could conclude that PaaR is also involved in branched-chain amino acid metabolism. It should be noted that the degree of transcriptional upregulation of bkdA is lower than that of bkdR upon paaR deletion (Fig. 4). As described before, bkdR exerted a negative effect on bkdABC transcription (16); thus, it could be proposed that the relatively lower transcriptional upregulation of bkdA in the ΔpaaR mutant may be the result of the combined role of PaaR and BkdR.
Effects of paaR deletion on the transcription of the paa genes. RNA samples were prepared from the fermentation cultures of the parental strain HCCB10218 and the ΔpaaR mutant (collected at 48 h). The transcription levels for each tested gene were normalized to the internal control, the hrdB gene. The values for the transcription of each gene in the parental strain were arbitrarily assigned as 1. The relative transcription levels are averages of three independent biological replicates, and error bars represent the standard deviations.
Subsequently, we deleted the paaABCDE operon based on the ΔpaaR mutant, generating a ΔpaaR ΔpaaABCDE double mutant, to verify that the reduced PI production in the ΔpaaR mutant is due to derepression of paa transcription. Pristinamycin production was measured by bioassay and HPLC analysis. Three strains, including the parental strain HCCB10218, the ΔpaaR mutant, and the ΔpaaR ΔpaaABCDE double mutant, were grown in fermentation medium, and samples were collected at 48 and 72 h. As expected, the results from the bioassay (fermentation samples collected at 48 h) revealed that unlike the ΔpaaR mutant, which produced a small inhibition zone, the ΔpaaR ΔpaaABCDE double mutant and the parental strain produced comparable clear zones (Fig. 5A). Further HPLC analysis showed that in comparison with the ΔpaaR mutant, which produced only PI levels that were about 30% of those of the parental strain, the ΔpaaR ΔpaaABCDE double mutant produced nearly the same amounts of PIa as those of the parental strain at the tested time points (Fig. 5B). Interestingly, compared with both the parental strain and the ΔpaaR mutant, the ΔpaaR ΔpaaABCDE double mutant produced 20% lower PIIa levels. Nevertheless, these results demonstrated that reduced PI production in the ΔpaaR mutant is due to the enhanced expression of paa genes encoding the PAA degradation pathway.
Effects of deletion of both paaR and paaABCDE on pristinamycin production. (A) Bioassay of pristinamycin production using B. subtilis as the indicator strain. Fermentation samples were collected at 48 h. (B) HPLC analysis of pristinamycin production. PIa and PIIa are the main components of PI and PII, respectively. Fermentation samples were collected at two time points (48 and 72 h). Error bars represent the standard deviations from three biological replicates. Three strains, including the parental strain HCCB10218, the ΔpaaR mutant, and the ΔpaaR ΔpaaABCDE double mutant (ΔpaaR/ABCDE), were tested.
Addition of l-Phg to the ΔpaaR mutant partially restores PI formation.To confirm that the reduced PI formation is due to the decreased l-Phg supply, we performed supplementation experiments by adding l-Phg to the fermentation medium of the ΔpaaR mutant and the parental strain HCCB10218. These two strains were incubated in seed medium for 44 to 48 h and were subsequently transferred into 25 ml of fermentation medium supplemented with 10 mg l-Phg (Sigma-Aldrich, MO, USA). The same two strains without the addition of l-Phg were tested as negative controls. After incubation in fermentation medium for 48 and 72 h, fermentation samples (0.5 ml) were collected and pristinamycin production was analyzed by both bioassay and HPLC. Bioassay showed that the addition of l-Phg to the ΔpaaR mutant resulted in a much larger clear zone than that of the mutant without the addition of l-Phg (samples collected at 48 h). For the parental strain, intriguingly, we observed that addition of l-Phg resulted in a slightly smaller inhibition zone than that without l-Phg (Fig. 6A). Further HPLC analysis revealed that, as shown in Fig. 6B, l-Phg addition to the ΔpaaR mutant could partially restore PI production at the tested time points; PI production was increased about 2-fold (from 15 to 30 mg/liter). However, to our surprise, l-Phg addition to the mutant resulted in enhanced PII production by around 20%. For the parental strain, addition of l-Phg led to an intriguingly enhanced PII but reduced PI production at the tested time points (Fig. 6B). Since the synergistic bacteriostatic activity of PI and PII is possibly dependent on the ratios of these two components (17), we speculated that a higher ratio of PII to PI may explain why the parental strain supplemented with l-Phg (PII:PI ratio ≈ 10:1) produced a slightly smaller clear zone than that without l-Phg (PII:PI ratio ≈ 4:1) (Fig. 6A).
Effects of l-Phg addition on pristinamycin production of the ΔpaaR mutant. (A) Bioassay of pristinamycin production using B. subtilis as the indicator strain. Fermentation samples with or without l-Phg supplementation were collected at 48 h. (B) HPLC analysis of pristinamycin production. PIa and PIIa are the main components of PI and PII, respectively. Fermentation samples were harvested at two time points (48 and 72 h). Error bars represent the standard deviations from three biological replicates. Two strains (the parental strain HCCB10218 and the ΔpaaR mutant) were tested under the condition of fermentation medium with or without the addition of l-Phg (10 mg/25 ml medium).
PaaR represses the transcription of paa genes directly.A previous report has revealed that PaaR from Actinobacteria recognizes a conserved, perfect palindromic motif (5′-ACCGA-n4-TCGGT-3′); moreover, in three Streptomyces strains, including S. coelicolor, S. avermitilis, and S. griseus, the identified PaaR-binding motifs were predicted to be located in the promoter regions of the paa genes as well as in the intergenic region of bkdA-bkdR (11). In this study, to determine whether it is the same case in S. pristinaespiralis, we searched for this signature sequence in the intergenic regions of paaH-paaZ, paaA-SSDG_03038, paaK-SSDG_07468, and bkdA-bkdR, which harbors the promoter regions of paaH and paaZ, paaA, paaK, and bkdA and bkdR, respectively. In addition, according to the genetic organization of paaI and its neighboring genes, we speculated that paaI is cotranscribed with the upstream four genes (from SSDG_03653 to SSDG_3650). Thus, the intergenic region of SSDG_03653-SSDG_05964 was also examined. The analysis revealed that the intergenic regions of paaH-paaZ, paaA-SSDG_03038, paaK-SSDG_07468, and bkdA-bkdR have the same conserved, perfect palindromic motif (5′-ACCGA-n4-TCGGT-3′) as that identified in C. glutamicum (11). Interestingly, in the former two regions (paaH-paaZ, paaA-SSDG_03038), we also found another imperfect palindromic motif (5′-AACGA-n4-TCGGT-3′), which overlaps with the perfect motif (Fig. 7A). No such motifs were detected in the intergenic region of SSDG_03653-SSDG_05964, which may explain why paaR deletion has no effect on paaI transcription. By using comparative genomic analysis, two such conserved PaaR-binding motifs were detected in the intergenic regions of the paa genes and between bdkA and bkdR in the genomes of six other Streptomyces strains, whose genome sequences are available online (http://streptomyces.org.uk/) (see Table S4 in the supplemental material), further confirming that PaaR-mediated control of these genes is common in Streptomyces strains.
Nucleotide sequences of the respective intergenic regions of paaH-paaZ, paaA-SSDG_03038, paaK-SSDG_07468, and bkdA-bkdR and their binding with purified PaaR protein. (A) Nucleotide sequences of the respective intergenic regions of paaH-paaZ, paaA-SSDG_03038, paaK-SSDG_07468, and bkdA-bkdR. The perfect conserved PaaR-binding motif is shaded, and the imperfect motif is indicated with broken lines. The transcriptional start sites (TSSs) are indicated by bent arrows. The putative −35 and −10 regions of the promoters of bkdA, paaZ, paaA, and paaK are underlined. The translational start and stop codons are marked by boxes. The primers used for amplification of EMSA probes are also underlined. (B) EMSA analysis. The concentrations of purified His6-PaaR protein (PaaR) used (nanomolar) in the assays are as indicated. We also performed competition assays with the addition of 200-fold specific (unlabeled each probe, indicated by S) and nonspecific (sperm DNA, indicated by N). The probe carrying the intergenic region of SSDG_03653-SSDG_05964 (containing the possible paaI promoter region) was tested as a negative control. (C) EMSA analysis in the presence of PA-CoA, PAA, or acetyl-CoA. The same concentration (200 nM) of purified His6-PaaR protein (PaaR) was used. PA-CoA, PAA, or acetyl-CoA was added at a concentration of 0, 10, 25, 50, 100, 200, or 400 μM. Free DNA probes and PaaR-probe DNA complexes are indicated by arrows.
Subsequently, EMSA analysis was performed with purified PaaR protein to investigate whether PaaR could specifically bind to these intergenic regions containing the signature sequences. In total, four probes containing the intergenic regions of paaH-paaZ, paaA-SSDG_03038, paaK-SSDG_07468, and bkdA-bkdR, respectively, were checked in the EMSA analysis(see Table S3 in the supplemental material). The intergenic region of SSDG_03653-SSDG_05964 that contains no putative PaaR-binding motif was tested as a negative control. As expected, purified PaaR protein could interact specifically with four tested probes (Fig. 7B) but not with the negative control, suggesting that the transcription of these paa genes was directly regulated by PaaR. To test whether PaaR was also involved in regulation of the transcription of two functionally unidentified genes, SSDG_03038 and SSDG_07468, which are transcribed divergently from paaA and paaK, respectively, qPCR was performed and revealed that PaaR repressed the transcription of SSDG_03038 as well and has no effect on SSDG_07468 expression (Fig. 4).
Furthermore, to determine whether as in C. glutamicum, the PaaR effector ligand is PA-CoA (11), we performed EMSA analysis of PaaR and the probe paaA-SSDG_03038p in the presence of PA-CoA (at concentrations of 0, 10, 25, 50, 100, 200, and 400 μM). PAA and acetyl-CoA at the same concentrations were tested as negative controls. As shown in Fig. 7C, we observed that the addition of increasing concentrations of PA-CoA could eventually abolish the binding activity of PaaR for paaA-SSDG_03038p; however, no effects were detected after adding either PAA or acetyl-CoA under the same EMSA conditions. These results confirmed that PA-CoA acts as the PaaR effector ligand.
Finally, to determine the positions of the PaaR-binding sites with respect to paa gene promoters, we performed 5′-RACE analysis to identify the putative transcriptional start sites (TSSs) of these paa genes that are directly regulated by PaaR, including paaH, paaZ, paaA, and paaK. RNA samples were isolated from fermentation cultures of the ΔpaaR mutant collected at 30 h. We successfully determined the TSSs of three paa genes (paaA, paaK, and paaZ), and the putative promoter structures were predicted, as shown in Fig. 7A and in Fig. S3 in the supplemental material. Additionally, according to the identified TSS of bkdA2 in S. coelicolor (a homologue of bkdA in S. pristinaespiralis) (16), we found the TSS and −10 and −35 regions of the bkdA promoter. Based on these results, we observed that the PaaR-binding sites are located either downstream of −10 regions or between −35 and −10 regions of the promoters of its target genes, which would hinder the binding of RNA polymerase to the promoter regions and thereby inhibit gene transcription.
DISCUSSION
In this study, a TetR family regulator, PaaR, was identified as being involved in the regulation of pristinamycin I (PI) by affecting the supply of one of seven amino acid precursors, l-Phg, in S. pristinaespiralis. A possible model for PaaR-mediated regulation of PI production was proposed as follows. Briefly, PaaR serves as a repressor in the expression of paa genes encoding the PAA degradation pathway. Its inactivation results in enhanced paa transcription, which would result in more consumption of PA-CoA, the common intermediate for the l-Phg biosynthetic pathway and PAA catabolic pathway (6, 11) (Fig. 1). Therefore, in the ΔpaaR mutant, the metabolic flux of PA-CoA into the l-Phg biosynthetic pathway was reduced, leading to a lower level of l-Phg and, accordingly, reduced PI biosynthesis. To our knowledge, this is the first report describing the interplay between the PAA degradation pathway and antibiotic biosynthesis in Streptomyces strains. As previously reported, the PAA catabolic pathway and its regulation by PaaR are widespread in antibiotic-producing actinomycetes (11). In addition, in bacteria, a substantial part of aromatic compounds (such as l-phenylalanine) is catabolized via the PAA pathway (18). Therefore, we speculated that PaaR-dependent regulation of antibiotic biosynthesis (particularly for antibiotics using aromatic compounds as the building precursors) may commonly exist.
Interestingly, although inactivation of paaABCDE based on the ΔpaaR mutant could restore PI titers comparable to those of the parental strain, the resulting ΔpaaR ΔpaaABCDE double mutant produced approximately 20% lower PIIa titers than the parental strain (Fig. 5). Given that a substantial part of aromatic compounds is degraded by the PAA catabolic pathway to acetyl-CoA (18), deletion of paaABCDE may lead to reduced acetyl-CoA formation and accordingly a lower level of malonyl-CoA, which serves as one of the extender units for PII biosynthesis (5). The decreased malonyl-CoA formation may account for the reduced PII biosynthesis upon paaABCDE deletion. Now, another question arises: why did the ΔpaaR mutant, with massively enhanced expression of the paa genes, still produce the same PII levels as those of the parental strain? We speculated that the expression of paa genes in the parental strain is sufficient for the degradation of aromatic compounds. Therefore, enhanced paa gene expression in the ΔpaaR mutant would not result in a significantly increase in malonyl-CoA level and therefore has little effect on PII biosynthesis.
Here, we found that although addition of l-Phg to the ΔpaaR mutant could partially rescue PI production, surprisingly, l-Phg supplementation led to enhanced PII titers in the mutant (Fig. 6). Furthermore, interestingly, l-Phg addition to the parental strain led to a significantly enhanced PII production but reduced PI production. As we know, l-Phg formation probably initiates from phenylpyruvate (PP) (Fig. 1) (5, 6), a compound from the shikimate pathway, which is the common pathway for the biosynthesis of aromatic compounds in bacteria and plants, such as l-tryptophan, l-tyrosine, and l-phenylalanine. In bacteria, the activity of key enzymes of the shikimate pathway, including prephenate dehydratase (PDT), chorismate mutase (CM), and 3-deoxy-d-arabinoheptulosonate-7-phosphate (DAHP) synthase, is under strict regulation via feedback inhibition. For instance, all three of these aromatic amino acids could repress the activity of DAHP synthase, which is responsible for the formation of DAHP (the first intermediate of the shikimate pathway) from erythose 4-phosphate (an intermediate from the pentose phosphate pathway [PPP]) and phosphoenolpyruvate (an intermediate from the glycolytic pathway) (19). This would reduce the metabolic flux to the biosynthesis of aromatic amino acids; on the other hand, it would enhance the metabolic flux to the PPP pathway for a greater NADPH supply as well as to the glycolytic pathway for the formation of acetyl-CoA and then malonyl-CoA. If this is also the case for S. pristinaespiralis, the addition of a high concentration of l-Phg (10 mg/25 ml) may result in a similar end product inhibition (feedback inhibition) and thereby lead to greater malonyl-CoA formation and, finally, enhanced PII biosynthesis. In addition, the biosynthetic pathway of DMAPA, one of the PI amino acid precursors, starts from chorismic acid, which is also an intermediate from the shikimate pathway (5). The end product repression would result in less carbon flow toward the DMAPA biosynthetic pathway. This may explain the findings that addition of l-Phg to the parental strain resulted in reduced PI biosynthesis and that its addition to the ΔpaaR mutant could lead to only partial restoration of PI titers.
ACKNOWLEDGMENTS
This study was funded by the National High Technology Research and Development Program of China (2012AA022107), the National Natural Science Foundation of China (31121001, 31370081, and 31430004), and the National Basic Research Program of China (2012CB721103). We also acknowledge the support of the SA-SIBS scholarship program.
FOOTNOTES
- Received 21 January 2015.
- Accepted 30 March 2015.
- Accepted manuscript posted online 13 April 2015.
- Address correspondence to Yinhua Lu, yhlu{at}sibs.ac.cn.
Citation Zhao Y, Feng R, Zheng G, Tian J, Ruan L, Ge M, Jiang W, Lu Y. 2015. Involvement of the TetR-type regulator PaaR in the regulation of pristinamycin I biosynthesis through an effect on precursor supply in Streptomyces pristinaespiralis. J Bacteriol 197:2062–2071. doi:10.1128/JB.00045-15.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.00045-15.
REFERENCES
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
