Redox-sensing regulator Rex regulates aerobic metabolism, morphological differentiation, and avermectin production in Streptomyces avermitilis

The regulatory role of redox-sensing regulator Rex was investigated in Streptomyces avermitilis. Eleven genes/operons were demonstrated to be directly regulated by Rex; these genes/operons are involved in aerobic metabolism, morphological differentiation, and secondary metabolism. Rex represses transcription of target genes/operons by binding to Rex operator (ROP) sequences in the promoter regions. NADH reduces DNA-binding activity of Rex to target promoters, while NAD+ competitively binds to Rex and modulates its DNA-binding activity. Rex plays an essential regulatory role in aerobic metabolism by controlling expression of the respiratory genes atpIBEFHAGDC, cydA1B1CD, nuoA1-N1, rex-hemAC1DB, hppA, and ndh2. Rex also regulates morphological differentiation by repressing expression of wblE, which encodes a putative WhiB-family transcriptional regulator. A rex-deletion mutant (Drex) showed higher avermectin production than the wild-type strain ATCC31267, and was more tolerant of oxygen limitation conditions in regard to avermectin production.

Scientific RepoRts | 7:44567 | DOI: 10.1038/srep44567 Even though Rex was first characterized in S. coelicolor and its regulatory mechanism has been extensively studied, few target operons/genes of Rex in Streptomyces have been confirmed 4 , and the overall regulatory function of Rex in this genus remains to be elucidated. S. avermitilis is an important species used for industrial production of avermectins, a group of anthelmintic antibiotics widely used in the medical, veterinary, and agricultural fields 11 . We investigated the regulatory role of Rex in the expression of operons/genes involved in aerobic metabolism, morphology, and secondary metabolism of S. avermitilis. Our findings have potential application to novel genetic engineering strategies for high antibiotic-producing strains and hypoxia-tolerating strains of this genus.

Results
Expression of atpIBEFHAGDC, cydA1B1CD, nuoA1-N1, and rex-hemAC1DB is negatively regulated by Rex. The rex gene is conserved within the genus Streptomyces and is cotranscribed with the heme synthesis genes hemACD 4 . To evaluate the regulatory role of Rex in S. avermitilis, we constructed a rex-deletion mutant (termed Drex) by homologous recombination in wild-type strain ATCC31267. rex deletion had no effect on growth in liquid fermentation medium (Fig. 1).
Expression of these genes differed greatly when cells were static-cultured following 3 days' culture on a rotary shaker (250 rpm). Transcription level of cydA1 under oxygen limitation condition in ATCC31267 increased steadily during 60 min, whereas the level in Drex increased to a maximal value during the first 30 min, then gradually declined during the subsequent 30 min (Fig. 2). These findings suggest that induction of cydA1 under oxygen limitation condition is mediated by Rex. Expression of nuoA1 and hemA under oxygen limitation increased slightly in the first 10 min, then declined during the subsequent 50 min, in both ATCC31267 and Drex. In contrast, expression of atpI under oxygen limitation declined steadily during 60 min in both ATCC31267 and Drex (Fig. 2). Transcription levels of cydA1, nuoA1, hemA, and atpI were consistently higher for Drex than for ATCC31267 under equivalent treatments, confirming that these genes are negatively regulated by Rex.
C-terminal His 6 -tagged Rex fusion protein was overexpressed in E. coli and purified for DNA binding analysis. EMSAs were performed to evaluate interactions between Rex and the promoters in vitro. Rex-His 6 bound to the promoter regions of atpIBEFHAGDC, cydA1B1CD, nuoA1-N1, and rex-hemAC1DB operons (Fig. 3A). ChIP assays were performed to assess interactions in vivo. ATCC31267 and Drex cells were treated with formaldehyde at days 2 and 6 to cross-link Rex to its DNA targets. Cross-linked DNA was extracted, fragmented by sonication, and immunoprecipitated by anti-Rex antibodies for screening of Rex-bound DNA fragments. In comparison to control hrdB promoter, PCR products of rex, cydA1, atpI, and nuoA1 promoter regions were selectively enriched from immunoprecipitated DNA of ATCC31267, whereas no such PCR bands were amplified from immunoprecipitated DNA of Drex (Fig. 3B). Results of EMSAs and ChIP assays revealed that Rex binds specifically to the promoter regions of atpIBEFHAGDC, cydA1B1CD, nuoA1-N1, and rex-hemAC1DB operons.
Determination of Rex operator (ROP) sequences on promoter regions of atpIBEFHAGDC, cydA1B1CD, nuoA1-N1, and rex-hemAC1DB. Rex binding sequences in 5′ -end fluorescein-labeled promoter regions of the above operons were determined by DNase I footprinting analysis. One protected region was detected in the rex promoter region in the presence of 1.2 or 2.4 μ Μ Rex-His 6 . The region extends for 23 nucleotides from positions − 39 to − 17 relative to the transcriptional start site (TSS) of rex. A consecutive ROP site (5′ -TTGTGCACGCGTTCACAA-3′ ) was found in the protected region; the site is located between − 35 region and − 10 region and encompasses − 35 region (Fig. S2A). A 28-nt protected region (positions − 3 to + 25 relative to TSS) was detected in the cydA1 promoter region. One ROP site (5′ -ATGTGAACGCGTTCACAA-3′ ) was found in the protected region downstream from TSS. A half-site ROP (5′ -TTGTGAA-3′ ) was also found in the protected region; it is located upstream from the ROP site and encompasses TSS (Fig. S2B). EMSA revealed two retarded bands between Rex and the cydA1 promoter region (Fig. 3A), suggesting that Rex can interact with the half-site ROP. A 29-nt protected region (positions − 50 to − 22 relative to TSS) containing a ROP site (5′ -TTGTGATACGGTTCACGA-3′ ) was detected in the atpI promoter region (Fig. S2C). Rex-His 6 protected a 27-nt region extending from positions − 42 to − 16 relative to TSS of nuoA1, which contains a ROP site (5′ -TTGTGACCTGCTTCACAT-3′ ) (Fig. S2D). ROP in the nuoA1 and atpI promoter regions is located between − 35 region and − 10 region, and encompasses − 35 region. These findings suggest that Rex blocks attachment of RNA polymerase to the promoters or inhibits the progress of RNA polymerase by binding to ROP in or  (B) ChIP assay analysis of Rex binding to promoter regions of rex, cydA1, atpI, nuoA1, and wblE in vivo. Rex-DNA complexes were immunoprecipitated by anti-Rex antibodies from formaldehyde-treated ATCC31267 and Drex cells. DNAs used for PCR were: total DNA prior to immunoprecipitation (positive control: lanes "+ "), immunoprecipitated DNA (experimental sample: lanes "S"), and DNA without antibody (negative control: lanes "− "). hrdB promoter region was used as control.
DNA-binding activity of Rex is modulated by NADH/NAD + ratio. In S. coelicolor, NADH at concentrations < 5 μ M inhibits DNA-binding activity of Rex, whereas 1 mM NAD + has no inhibitory effect. NAD + competes with NADH for Rex binding 4 . In B. subtilis and S. aureus, NAD + enhances binding of Rex to putative Rex-binding sites, while NADH competes with NAD + for Rex binding and reduces Rex activity 6,7 . We examined the effects of NAD + and NADH on DNA-binding activity of Rex to upstream regions of cydA1 in S. avermitilis. DNA-binding activity of Rex was reduced by addition of NADH, but not by NAD + concentrations up to 1 mM (Fig. 4A,B; Fig. S3). NADH and NAD + were added to EMSA binding buffer to assess the effect of NAD + / NADH ratio on DNA-binding activity of Rex in vitro. At NAD + concentration 0.2 mM, 5 μ M NADH was sufficient to dissociate the Rex-DNA complex (Fig. 4C). At NAD + concentration 1 mM, dissociation of DNA-Rex complex required 25 μ M NADH, suggesting that Rex-binding activity was recovered by addition of increasing amounts of NAD + (Fig. 4D). These findings indicate that NAD + and NADH bind competitively to Rex and modulate its DNA-binding activity. These findings also imply that Rex exploits the similar regulatory mechanism in Streptomyces.
Rex regulates morphological differentiation. In comparison to ATCC31267, Drex showed delayed morphogenesis on SFM agar at day 2, when aerial mycelium was initiated. Spore formation at day 6 did not differ notably between the two strains. Morphogenesis of the Drex complementation strain was similar to that of ATCC31267 (Fig. 5), indicating that the delayed morphogenesis was due solely to rex deletion. The promoter region of wblE in S. avermitilis contains a putative Rex-binding motif ( Table 1). wblE encodes a putative WhiB-family transcriptional regulator, which may be involved in morphological differentiation 12,13 . qRT-PCR analysis revealed notable increases of wblE transcription level in Drex. Levels under oxygen limitation condition declined gradually during 60 min for both ATCC31267 and Drex, and were consistently higher for Drex than for ATCC31267 (Fig. 2). EMSAs showed that Rex-His 6 bound to the wblE promoter region in vitro (Fig. 3A). In in vivo ChIP assays, PCR product of the wblE promoter region was selectively enriched from immunoprecipitated DNA of ATCC31267, whereas no such PCR band was amplified from immunoprecipitated DNA of Drex (Fig. 3B). These findings indicate that wblE is negatively regulated by Rex. Rex binding sequence in the wblE promoter region was determined by DNase I footprinting analysis. A 28-nt region protected by Rex-His 6 was detected, extending from positions + 108 to + 135 relative to TSS of wblE (Fig. S4). The protected region contains a consecutive ROP site (5′ -TCGTGAAAGCGTTCACAT-3′ ) and a half-site ROP (5′ -TTCACAA-3′ ) located downstream of TSS. Rex may inhibit the progress of RNA polymerase by binding to ROP downstream of the wblE promoter, and thereby repress transcription.
To test the possibility that overexpression of wblE in Drex results in delayed morphogenesis, we attempted to delete wblE in S. avermitilis. However, this attempt was unsuccessful. wblE is evidently an essential gene in Streptomyces; an attempt to delete it in S. coelicolor was also unsuccessful 13 . When wblE was overexpressed in ATCC31267, the resulting strain had a phenotype similar to that of Drex (Fig. 5), suggesting that Rex regulates morphological differentiation through its effect on wblE expression.
Rex negatively regulates avermectin production. The overexpression of rex caused a decrease in avermectin production to 33% of ATCC31267 level. Drex had avermectin production ~3-fold higher than that of ATCC31267. The mycelial dry weight of Drex was similar to that of ATCC31267, indicating that the improved avermectin yield was not achieved by improved growth. In the Drex complementation strain, avermectin production was similar to that of ATCC31267 ( Fig. 1; Fig. 6A). These findings indicate that rex negatively regulates avermectin production in S. avermitilis. We also measured avermectin production under oxygen limitation condition. In ATCC31267, lower agitation speed (230 rpm; control speed was 250 rpm) resulted in a 20% reduction of avermectin production, and static culture for 2 h on day 5 resulted in a 35% reduction. In Drex, avermectin production was reduced by 3.8% and 25%, respectively, under the above two conditions (Fig. 6B). Thus, rex deletion resulted in increased tolerance of S. avermitilis to oxygen limitation in regard to avermectin production.
qRT-PCR analysis was performed to determine whether Rex regulates avermectin production at the transcriptional level. Drex showed significantly increased transcription levels of pathway-specific regulatory gene aveR and biosynthetic genes aveA1 and aveD, relative to ATCC31267. Oxygen limitation for 60 min reduced expression of these genes in ATCC31267; however, Drex showed lower fold repression, and a slight induction of aveA1 and aveD (Fig. 6C). EMSAs revealed that Rex-His 6 did not bind to the aveR promoter region or the aveD-A1 intergenic region (Fig. S5). Although no retarded band was observed when aveR promoter region was probed with Rex-His 6 protein, DNase I footprinting analysis showed one protected region extending for 15 nucleotides on the aveR coding strand in the presence of 10 or 15 μ Μ Rex-His 6 (Fig. 7). No consecutive ROP site was observed in the protected region; however, two adjacent half-site ROP (5′ -TCGTGAA-3′ and 5′ -TTGTGGA-3′ ) were found in the protected region and downstream region. Rex can evidently interact with the half-site ROP; however, because the interaction is weak and easily dissociated in vitro, EMSA did not reveal a clear shifted band.

Confirmation of putative Rex target genes. A genome-wide search of consensus motif
5′ -TTGTGAANNNNTTCACAA-3′ using the genome sequence of ATCC31267 revealed the presence of 36 motifs up to 350 bp upstream of predicted genes: 2 motifs with one mismatch, 10 motifs with two mismatches, and 24 motifs with three mismatches. Our previous experiments showed that wblE, cydA1B1CD, rex-hemAC1DB, atpIBEFHAGDC, and nuoA1-N1 are directly controlled by Rex. To investigate whether Rex binds to promoter regions of other putative target genes, we selected 16 genes with predicted gene function for EMSAs (Table 1). Of these, Rex bound to the probes of hppA (encodes an inorganic H + pyrophosphatase), ndh2 (encodes a NADH dehydrogenase [complex I]), echA7 (encodes an enoyl-CoA hydratase), ectABC (encodes ectoine biosynthesis enzymes), SAV828 (encodes a rhamnosidase), and SAV2652 (encodes a regulatory protein). Probes whose binding motif had one or two mismatches showed higher affinity than probes whose binding motif had three mismatches (Fig. 3A, Fig. 8). These findings demonstrated that 5′ -TTGTGAANNNNTTCACAA-3′ is the consensus motif of Rex in S. avermitilis.

Discussion
Results of this study show that Rex in S. avermitilis acts as a repressor of aerobic metabolism, morphological differentiation, and secondary metabolism (summarized schematically in Fig. 9). Results of EMSAs demonstrated that at least 11 genes/operons are directly regulated by Rex. Among these, atpIBEFHAGDC, cydA1B1CD, nuoA1-N1, and rex-hemAC1DB operons encode key components of the electron transfer chain and play crucial roles in aerobic metabolism [14][15][16][17] . hppA encodes a putative pyrophosphate-energized proton pump that converts energy from pyrophosphate hydrolysis into active H + transport across the plasma membrane 18 . ndh2 encodes a NADH dehydrogenase involved in NAD + regeneration 19,20 . echA7 encodes an enoyl-CoA hydratase that catalyzes the second step of the β -oxidation pathway of fatty acid metabolism 21 . SAV828 encodes a rhamnosidase that hydrolyzes L-rhamnose from L-rhamnoside 22 . Under oxygen limitation condition, the increase of intracellular NADH/NAD + ratio in S. avermitilis dissociates binding of Rex from its target binding sites and derepresses its target genes/operons, and upregulation of cydA1B1CD, nuoA1-N1, rex-hemAC1DB, ndh2, and hppA increases oxygen utilization, NAD + regeneration, and ATP synthesis (Fig. 9). On the other hand, expression of atpIBE-FHAGDC in ATCC31267 and Drex is downregulated by oxygen limitation, suggesting that this operon is also directly controlled by regulators other than Rex. The F 0 F 1 -ATPase operon in Corynebacterium glutamicum is regulated by ECF σ H 23 . A sigH homolog is present in Streptomyces; whether it regulates atpIBEFHAGDC expression remains to be tested.
WhiB-like family transcription factors are widely present in actinomycetes, but not found in other bacterial orders. WhiB was first identified as a small transcription factor-like protein essential for sporulation in S. coelicolor 24 . Genome sequencing revealed that Streptomyces species have multiple whiB-like genes (designated "wbl"). Eleven wbl genes (including whiB and whiD) have been identified in S. coelicolor 13 . Among these, wblA, whiB, and whiD are essential for sporulation, and WblA also negatively regulates antibiotic biosynthesis in Streptomyces 13,25-27 . Other wbl genes are not involved in morphological development, with the exception of wblE. Fowler-Goldsworthy et al. 13 reported that wblE could not be deleted in various strains of S. coelicolor, and we made a similar observation in S. avermitilis. Thus, wblE appears to be essential in this genus. The homolog of wblE in Mycobacterium tuberculosis is whiB1, which encodes an essential transcription factor in response to nitric oxide exposure 28 . We demonstrated that wblE is directly negatively regulated by Rex, and that wblE overexpression results in delayed morphogenesis similar to that of Drex. Expression of wblE, like that of atpIBEFHAGDC, is downregulated by oxygen limitation in both ATCC31267 and Drex, suggesting that (i) wblE is jointly regulated by Rex and some other regulator, or (ii) wblE itself responds to low oxygen concentration via its own redox-sensitive [4Fe-4S] cluster. Under oxygen limitation condition, wblE expression in Streptomyces is downregulated, with consequent stimulation of sporulation and production of a large number of spores to maintain viability under conditions of little or no oxygen. Another Rex target, ectABC, encodes enzymes for biosynthesis of ectoine (a compatible solute) that serves as an osmolyte and promotes survival under osmotic or temperature stress 29 . By regulating ectABC transcription, Rex facilitates ectoine biosynthesis to enhance viability under these types of stress.
In Drex, expression of regulatory gene aveR and biosynthetic genes, and avermectin production, were notably increased. Although EMSA showed no clearly retarded band between aveR promoter region probe and Rex-His 6 , DNase I footprinting analysis revealed one 15-nt protected region consisting of two adjacent half-site ROP on the coding strand of aveR by Rex-His 6 . Thus, Rex may directly regulate aveR expression by interacting with the half-site ROP in the aveR promoter region. Expression of electron transfer chain components was enhanced in Drex, thus promoting aerobic respiration rate, ATP production, and secondary metabolism. The notable increase of atpIBEFHAGDC, cydA1B1CD, nuoA1-N1, and rex-hemAC1DB expression in Drex enhanced the tolerance of cells to oxygen limitation. The findings described here provide a basis for construction of new Streptomyces strains with high antibiotic production and hypoxia tolerance.

Materials and Methods
Bacterial strains and growth conditions. The S. avermitilis strains used were ATCC31267 (wild-type), Drex (rex-deletion strain), Drex-C (rex-deletion complementary strain harboring plasmid pSET-rex), and Orex (ATCC31267 harboring rex overexpressing plasmid pKC-rex). E. coli strains JM109 and BL21 (DE3) were used for routine cloning and protein expression, respectively. YMS medium and SFM medium were used for sporulation and phenotype studies 30,31 . Culture conditions for mycelial growth, protoplast preparation, and regeneration of S. avermitilis were as described previously 30 . Seed medium and fermentation medium FM-I were used for avermectin production and for RNA isolation, and soluble fermentation medium FM-II was used for ChIP analysis 32 .
Gene deletion, complementation, and overexpression. A rex (SAV4738) gene deletion mutant was generated through targeted gene deletion mediated by homologous recombination. A 566-bp fragment upstream of rex (position − 460 to + 87 from start codon) was amplified by primers rex-up-Fw and rex-up-Rev, and a 579-bp fragment downstream of rex (position +539 to +1098) was amplified by primers rex-dw-Fw and rex-dw-Rev, using ATCC31267 genomic DNA as template (Table S1; Fig. 1). The two fragments, after recovery, were digested respectively by BamHI/HindIII and BamHI/EcoRI, and ligated together into EcoRI/HindIII-digested pKC1139 33 to produce rex-deletion vector pKCD-rex. pKCD-rex was introduced into ATCC31267 protoplasts. Double-crossover recombinant strains were selected as described previously 34,35 . The rex-deletion mutant (termed Drex) was confirmed by PCR using one pair of external primers (rex-V-Fw/rex-V-Rev) and one pair of internal primers (rex-V2-Fw/rex-V2-Rev) (Table S1; Fig. 1). Use of the external primers yielded a 1.3-kb band from Drex and a 1.8-kb band from ATCC31267. Use of the internal primers yielded a 225-bp band from ATCC31267 and no band from Drex (data not shown).
A 1038-bp DNA fragment carrying the rex ORF and its putative promoter was amplified by PCR using primers rex-E-Fw and rex-E-Rev (Suppl. Table 1), and then ligated into EcoRI/XbaI-digested pSET152 or pKC1139 to produce vector pSET-rex or pKC-rex. For complementation analysis of Drex, pSET-rex was transformed into Drex protoplasts. For overexpression of Rex, pKC-rex was introduced into ATCC31267 protoplasts.
RNA extraction and qRT-PCR analysis. RNA was isolated using Trizol reagent (Tiangen; China) from S. avermitilis mycelia grown in FM-I as described previously 32 . Transcription levels of various genes were determined by qRT-PCR using the primer pairs listed in Table S1. An RNA sample without prior reverse transcription was used as negative control to rule out chromosomal DNA contamination. hrdB gene (SAV2444) was used as internal control.
Chromatin Immunoprecipitation (ChIP) assay. ChIP assay was performed as described previously 36 .
In brief, S. avermitilis cultures grown in FM-II for 2 or 6 days were fixed in cross-linking buffer (0.4 M sucrose, 10 mM Tris·Cl [pH 8.0], 1 mM EDTA) containing 1% formaldehyde for 20 min at 28 °C. ChIP was performed using anti-Rex antibody. After DNA extraction, pellets were washed with 70% ethanol and resuspended in 50 μ l Tris·EDTA buffer. 1 μ l DNA solution was subjected to PCR using the primer sets listed in Table S1. Overexpression and purification of Rex-His 6 . The rex coding region was amplified by PCR using primers His-rex-Fw and His-rex-Rev. The purified fragment was cut with NcoI/HindIII and cloned into NcoI/HindIII-digested pET28a (+ ) to generate expression plasmid pET-rex. pET-rex was introduced into E. coli BL21 (DE3) for overexpression of C-terminal His 6 -tagged Rex. Rex-His 6 was induced by 0.2 mM IPTG at 37 °C and purified from whole-cell lysate by Ni-NTA agarose chromatography (Bio-works; Sweden) according to the manufacturer's instructions.
Electrophoretic mobility gel shift assays (EMSAs). EMSAs were performed according to the manufacturer's instructions (DIG Gel Shift Kit, 2 nd Generation, Roche) as described previously 35 . DNA probes were obtained by PCR using the primers listed in Table S1, and labeled with Digoxigenin-11-ddUTP at the 3′ end using recombinant terminal transferase. DIG-labeled DNA probe was incubated with various quantities of Rex-His 6 for 30 min at 25 °C in a total volume of 20 μ l containing 1 μ g poly[d(I-C)]. Electrophoresis (5.0% native polyacrylamide gel; 0.5 × TBE as running buffer) was performed to separate protein-bound probes from free probes. DNA was electroblotted onto a positively charged nylon membrane, and retarded and unbound bands were detected by chemiluminescence and recorded on X-ray film.
DNase I footprinting assays. A fluorescent labeling procedure was used for DNase I footprinting assays 37 . DNA fragments were obtained by PCR using FAM-labeled primers (Table S1), and purified from agarose gel. Labeled DNA fragments (400 ng) and various quantities of Rex-His 6 were incubated in a 25-μ l volume for 30 min at 25 °C. DNase I digestion was performed for 40 sec at 37 °C, and terminated by addition of 10 μ l 0.2 M EDTA Fermentation and HPLC analysis of avermectin production. Fermentation of S. avermitilis strains and estimation of avermectins yields by HPLC analysis were performed as described previously 32 . Determination of transcriptional start sites. Transcriptional start sites (TSS) of rex and wblE were mapped by 5′ -RACE using a 5′ /3′ RACE Kit (2nd Generation, Roche). Total RNA was extracted from ATCC31267 grown in FM-I for 2 days. A gene-specific primer (sp1) was used to synthesize cDNA, and template RNA was degraded with RNase H. A homopolymeric A-tail was purified and added to the 3′ -end of cDNA using terminal transferase. Tailed cDNA was PCR amplified through 35 cycles with a specific nested primer (sp2) and an oligo (dT)-anchor primer (Table S1). PCR products were electrophoresed, purified using a DNA agarose gel recovery kit (BioTek; China), and sequenced.

Prediction of Rex putative targets.
To search for putative Rex target genes, Rex consensus motif 5′ -TTGTGAANNNNTTCACAA-3′ was used to scan the intergenic regions of the S. avermitilis genome using Virtual Footprint software 38 .