Characterization of γ-Butyrolactone Autoregulatory Signaling Gene Homologs in the Angucyclinone Polyketide WS5995B Producer Streptomyces acidiscabies^▿

Cloning of GBL signaling genes sabRAS in S. acidiscabies 84.104.

All oligonucleotide primers used in this study are presented in Table ​Table2.2. Consensus degenerate hybrid oligonucleotide primers (39) were selected using protein alignments of previously characterized GBL receptors. The oligonucleotide primer pair GBL-F/GBL-R (Table ​(Table2)2) was used to amplify GBL receptor gene fragments from S. acidiscabies 84.104 genomic DNA using “touchdown PCR” as previously described (13). Cloned amplification products were sequenced and used as probes for the identification of complete genes and flanking regions from gene libraries using colony hybridization methods (13). The nucleotide sequence of the sabRAS region was determined on both strands of a 7-kb KpnI insert in plasmid vector pUC19 (pGBLBP1) using a primer walking approach. Deduced GBL receptor and synthase protein sequences were aligned using MUSCLE, version 3.7. Alignments were refined using Gblocks, version 0.91b, and phylogenies were constructed with the maximum-likelihood method using PhyML, version 3.0. Phylograms were visualized and edited using TreeDyn, version 198.3 (9).

Mutagenesis and complementation of sabRAS genes.

The 7-kb KpnI fragment from pGBLBP1 was cloned into the KpnI site of plasmid vector pOJ260, and the resulting plasmid (pALE21) was introduced into E. coli strain BW25113 carrying plasmid pKD46, encoding bacteriophage lambda Red recombinase. Cultures carrying both of these plasmids were grown in LB medium containing apramycin and ampicillin. BW25113(pALE21/pKD46) was grown in SOB medium containing 1 mM l-arabinose to induce expression of Red recombinase, as previously described (8). Oligonucleotide primer pairs sabRdelF/sabRdelR, sabAdelF/sabAdelR, and sabSdelF/sabSdelR (Table ​(Table2)2) were selected to generate in-frame deletions of sabR, sabA, and sabS on plasmid pALE21, respectively. These primers were used to amplify the chloramphenicol acetyltransferase gene from template plasmid pKD3, with resulting products carrying 5′ ends with homology to sabR, sabA, or sabS in pALE21 (8, 11). Gel-purified PCR products were electroporated into washed suspensions of BW25113(pALE21/pKD46) as described previously (8), and transformed cells carrying mutagenized pALE21 were selected on LB agar containing chloramphenicol. Overnight cultures of resulting chloramphenicol-resistant colonies were grown in LB medium containing chloramphenicol, and plasmids were extracted from these cultures. Plasmids thus obtained were used to transform suspensions of competent E. coli DH5α cells to chloramphenicol resistance. Plasmids extracted from these transformants were used to transform BW25113(pCP20) to apramycin resistance in order to excise the chloramphenicol resistance gene as described previously (8).

pALE21 derivatives carrying deletions in sabR, sabA, and sabS were used to transform either E. coli strain S17-1 or the ET12567(pUZ8002) strain to apramycin resistance. Resulting transformants were used for intergeneric conjugal matings with S. acidiscabies 84.104 recipient strains as previously described (13), with the exception that mating mixtures were plated on 25 ml of modified ISP4 medium instead of AS-1 medium. Following overnight growth of mating mixtures at 30°C, transconjugants were selected by overlaying plates with 5 ml of soft nutrient agar containing nalidixic acid and apramycin sulfate (150 μg/ml each) to give final concentrations of 25 μg/ml in agar medium. Incubation of plates was continued at 30°C for 3 to 5 days until transconjugants were visible on agar surface. Transconjugant colonies were then transferred to ISP2 medium containing nalidixic acid and apramycin. Liquid cultures of transconjugants were grown nonselectively in tryptic soy broth for two or three serial transfers to allow for loss of integrated plasmid. PCR assays were used to screen apramycin-sensitive derivatives for loss of plasmid and retention of mutant alleles. These mutants were used to study the roles of GBL signaling genes in secondary metabolite biosynthesis and morphological differentiation on solid medium.

In order to verify linkage of mutant phenotypes with deleted GBL alleles, the primer pairs sabRcomplF/sabRcomplR, sabAcomplF/sabAcomplR, and sabScomplF/sabScomplR were used to amplify the wild-type sabR, sabA, and sabS open reading frames (ORFs), respectively, from S. acidiscabies 84.104 genomic DNA using high-fidelity Pfu polymerase (New England Biolabs). Amplification products were purified and digested with restriction enzymes HindIII and BglII. Plasmid pIJ86 was also digested with HindIII and BglII, and the ORFs encoding these GBL signaling genes were ligated into pIJ86 using T4 DNA ligase (New England Biolabs). The resulting plasmids, pIJ86::sabR, pIJ86::sabA, and pIJ86::sabS, harboring wild-type genes under the transcriptional control of the ermE* promoter were used to transform E. coli S17-1 or ET12567(pUZ8002), and conjugations were performed using these strains as donor and the ΔsabR, ΔsabA, or ΔsabS strain as recipient. Resulting transconjugants were selected on ISP4 medium amended with apramycin and nalidixic acid.

Analysis of WS5995B production.

S. acidiscabies 84.104 wild type and GBL deletion mutants were grown on ISP2 medium, and mycelial fragments were used to inoculate tryptic soy broth liquid medium (20 ml in 250-ml baffled Erlenmeyer flasks). Cultures were grown at 30°C and 180 rpm for approximately 20 h. Two milliliters of these cultures was used to inoculate 280 ml of SGM in 2.8-liter baffled Fernbach flasks. Culture medium was collected at various times, and cells were removed by vacuum filtration. Wet cell weight determinations of cultures were made, and filtrates were extracted using Strata-X solid-phase cartridges as described by the manufacturer (Phenomenex).

Extracted metabolites were dissolved in methanol and investigated by thin-layer chromatography (TLC) analysis on silica gel plates (250 μm; Whatman). UV and visible absorbance spectra were collected using a Beckman DU640B spectrophotometer. Quantitative high-performance liquid chromatography analysis was done using a Hewlett Packard 1090 liquid chromatograph equipped with diode-array detector and a Phenomenex C₁₈ Luna (5 μm; 150 by 4.6 mm) column. Detection wavelength range was from 200 to 600 nm. Solvent A was H₂O·H₃PO₄ at 99.9:0.1; solvent B was 100% CH₃CN. Samples of extracts were injected and chromatographed at a 1 ml/min flow rate. Extract components were eluted with a linear gradient of 25% B to 60% B over 21 min. Identification and quantification of WS5995B in culture filtrate extracts were accomplished using authentic WS5995B standard preparations (21).

Cloning, expression, and purification of recombinant SabR and SabS.

The sabR and sabS ORFs were amplified from S. acidiscabies 84.104 genomic DNA using high-fidelity Pfu polymerase and the oligonucleotide primer pairs sabRex-F/sabRex-R (for SabR expression) and sabSex-F/sabSex-R (for SabS expression) (Table ​(Table2).2). Amplification reaction products were purified and digested with restriction enzymes NdeI and XhoI. The expression vector pET26b (Novagen) was also digested with NdeI and XhoI, and sabR and sabS ORF restriction digestion products were ligated into the vector to produce recombinant protein expression constructs pET26b::sabR and pET26b::sabS. Competent cells of E. coli strain BL21(λDE3) were transformed with these plasmids, and transformants were selected on LB agar containing kanamycin. Transformants were grown in 500 ml of LB medium containing kanamycin at 30°C. When the culture reached an optical density at 600 nm of approximately 0.6, protein expression was induced with the addition of IPTG (isopropyl-β-d-thiogalactopyranoside) to a final concentration of 80 μM. Cultures were grown for an additional 6 h and then harvested by centrifugation. Cell pellets were washed once in buffer containing 20 mM sodium phosphate, 500 mM NaCl, and 20 mM imidazole, pH 7.4, and then resuspended in 1 to 2 ml of the same buffer containing 1 mM dithiothreitol (DTT) and 1 mM phenylmethylsulfonyl fluoride. Cell suspensions were disrupted by passage through a French pressure minicell at 16,000 lb/in². The cell lysate was centrifuged at 20,000 × g for 30 min, and the resulting soluble protein fraction was loaded onto a 1-ml HisGraviTrap column (GE Healthcare). Protein fractions were eluted with 20 mM sodium phosphate-500 mM NaCl, pH 7.4, containing increasing amounts of imidazole. Fractions were analyzed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (PAGE), and the fractions containing pure SabR or SabS (as judged by electrophoretic homogeneity) were pooled and desalted on PD-10 desalting columns (GE Healthcare). Proteins obtained in this manner were stored at −20°C or −70°C in 50 mM Tris-Cl, pH 7.6, 200 mM NaCl, 1 mM DTT, and 50% glycerol. Protein determinations were made using a Bradford assay with bovine serum albumin as the standard.

Electrophoretic mobility shift assays.

DNA fragments encompassing ARE sequences within the sabRAS region were amplified from template plasmid pGBLBP1. The oligonucleotide primer pair ARE-F/ARE-R was used to amplify a 208-bp DNA fragment encompassing the upstream region of sabS. ARE-F (40 pmol) was first end labeled in a separate reaction using [γ-³²P]ATP (6,000 Ci/mmol; PerkinElmer) and polynucleotide kinase (Roche) according to the manufacturer's recommendations. Labeled primer products were separated from unincorporated nucleotide using Micro Bio-Spin columns (Bio-Rad) and used in PCR along with oligonucleotide primer ARE-R, plasmid template pGBLBP1, and high-fidelity Pfu polymerase. Amplified labeled PCR product was separated from excess primer using QiaQuick PCR spin columns (Qiagen). DNA binding assays contained 2.5 × 10³ cpm of labeled DNA, 50 ng of calf thymus DNA, 50 mM NaCl, 10 mM MgCl₂, 10 mM Tris-Cl, 1 mM DTT, and 50 μg/ml bovine serum albumin, pH 7.4. Various amounts of recombinant SabR or SabS were added to the reaction tubes. For competition experiments, unlabeled competitor duplex DNA corresponding to ARE sequence was prepared by briefly heating and annealing stoichiometric amounts of complementary oligonucleotides AREdupF/AREdupR and adding these to DNA binding reaction mixtures. Binding reactions (10 μl) were carried out for 20 min at 30°C, after which reaction products were electrophoresed through nondenaturing 5% polyacrylamide gels in 0.5× Tris-borate-EDTA buffer (40). Gels were dried under vacuum and exposed to a phosphorimager screen (Amersham Biosciences). Exposed screens were scanned using a Typhoon Trio+ Variable Mode Imager (Amersham Biosciences); band intensities of scanned screens were quantified using ImageQuant TL software (Amersham Biosciences). Experiments were repeated at least four times for each protein and ARE substrate combination.

RNA extraction and reverse transcriptase PCR assays.

Cultures of S. acidiscabies wild-type 84.104 and of sabR, sabA, and sabS mutants were grown in tryptic soy broth or SGM to mid-log phase, and cells were collected by centrifugation. Cell pellets were briefly homogenized using a tissue grinder and treated with lysozyme. Total RNA was extracted from cells using an RNeasy minikit following the manufacturer's guidelines (Qiagen). Following extraction, RNA samples were treated with RNase-free DNase I (New England Biolabs). Reverse transcription reactions were carried out using Superscript II reverse transcriptase (Invitrogen) and reverse transcriptase primers sabRRrt, sabARrt, and sabSRrt (Table ​(Table2).2). PCR amplification of DNase-treated samples prior to cDNA synthesis yielded no amplification products, indicating that nucleic acid PCR amplification products observed following reverse transcription were produced solely from cDNA templates derived from RNA. Following cDNA synthesis, reverse transcription reaction products were used for PCR amplification using the primer pairs sabRFrt/sabRRrt, sabAFrt/sabARrt, and sabSFrt/sabSRrt. PCR was carried out for 30 cycles, and amplification products were electrophoresed through 1% agarose gels in Tris-acetate-EDTA buffer. PCR amplification product band intensities on gels were imaged using a VersaDoc system (Bio-Rad) and quantified using Quantity One software. Normalization of amplification products from cDNA templates using GBL primers was done using the hrdB gene encoding the principal sigma factor of S. acidiscabies 84.104. hrdB was cloned from S. acidiscabies 84.104 genomic DNA using degenerate oligonucleotide primers designed from alignments of conserved regions of principal sigma factors from S. coelicolor (SCO5820), Streptomyces avermitilis (SAV_2444), and S. griseus (EMBL accession {"type":"entrez-nucleotide","attrs":{"text":"X75952","term_id":"440164","term_text":"X75952"}}X75952) (Table ​(Table2).2). The deduced amino acid sequence of the cloned S. acidiscabies 84.104 hrdB gene displayed 95% identity with S. coelicolor HrdB (not shown). The oligonucleotide primer pair hrdBFrt/hrdBRrt was used for amplification of hrdB cDNA from RNA samples.

Quantitative reverse transcription-PCR (qRT-PCR) was performed using an Applied Biosystems 7500 real-time PCR system and SYBR green Quantitative RT-PCR Kit with Moloney murine leukemia virus reverse transcriptase, RNase inhibitor, and JumpStart Taq DNA polymerase (Sigma). Specific primers used for amplification of sabS (sabSQRT) and hrdB (hrdBQRT) are given in Table ​Table2.2. Amplicon specificity was checked using qRT-PCR dissociation curve analysis. The relative increase in sabS expression in wild-type and sabR mutant strains was determined using the Relative Quantification Method (Applied Biosystems 7500 System). Briefly, threshold cycle (C_T) values were normalized to hrdB mRNA levels for RNA samples from each strain, and mean relative expression ratios were calculated using the ΔΔC_T method. Values given for the relative increase in expression represent the means of three independent experiments.

Characterization of sabR, sabA, and sabS mutants.

In order to study the roles of sabRAS genes in S. acidiscabies, in-frame deletion mutants of each gene were constructed using a modification of the bacteriophage lambda Red recombinase method developed by Datsenko and Wanner (8) and later adapted to Streptomyces (11). Our approach involved PCR amplification of the template plasmid pKD3 cat antibiotic resistance gene and electroporation mutagenesis of a derivative of the suicide plasmid vector pOJ260 (pALE21) which carries the cloned mutagenesis target genes and the origin of conjugal DNA transfer oriT (2). Previous conjugation experiments using S. acidiscabies utilized AS-1 medium for plating of conjugation mixtures (12, 13). More recent conjugation studies with S. acidiscabies in our lab have shown that improved results could be obtained using ISP4 medium amended with yeast extract and 40 mM MgCl₂. Experiments using this medium resulted in 10- to 30-fold increases in observed transconjugant frequency. Lower increases were seen with the addition of 10, 20, and 30 mM MgCl₂, and no difference in transconjugant numbers was observed between 40 mM and 50 mM MgCl₂ amendments (data not shown). Similar transconjugant frequencies were obtained using either E. coli conjugal donor strain S17-1 or ET12567(pUZ8002).

The wild-type strain and the ΔsabR, ΔsabA, and ΔsabS mutants were grown on ISP2 medium, SGM, and oatmeal agar medium to identify differences in colony morphology and/or in the appearance of pigmented secondary metabolites. SGM supports production of the yellow pigmented angucyclinone polyketide WS5995B (21), whereas oatmeal-based medium supports production of high levels of thaxtomins in S. acidiscabies (28). After growth in oatmeal broth medium, no differences in growth yields between the wild-type and mutant strains were observed (data not shown). Similarly, following extraction and silica gel and reverse-phase TLC analysis of oatmeal broth culture filtrates, no differences in thaxtomin production were seen between the wild-type and mutant strains (data not shown).

While no changes in thaxtomin production were seen between the wild type and the mutants, a comparison of the wild type and the ΔsabA and ΔsabS strains grown on solid SGM showed that these two mutants produced higher levels of a yellow diffusible pigment than the parent strain (Fig. ​(Fig.4A).4A). No significant differences in pigment production relative to the parent strain were observed in ΔsabR mutants on SGM or oatmeal agar medium (data not shown). To verify that the pigment overproduction phenotypes observed in sabA and sabS mutants were due to the mutant sabA and sabS alleles, the wild-type alleles of these genes were cloned into plasmid pIJ86 under the transcriptional control of the constitutive ermE* promoter (1). Normal wild-type levels of pigment production were seen in mutant transconjugants carrying the plasmids, demonstrating that pigment production phenotypes were due to mutations in sabA and sabS (Fig. 4B and C).

Open in a separate window

FIG. 4.

Phenotypic properties of sabR, sabA, and sabS mutants. (A) Pigment production characteristics of sabR, sabA, and sabS deletion mutants on SGM. wt, wild-type strain S. acidiscabies 84.104. (B) Complementation of sabS deletion mutant on SGM with plasmid pIJ86::sabS. (C) Complementation of sabA deletion mutant on SGM with plasmid pIJ86::sabA. (D) Complementation of conditional morphological defect of sabS mutant on ISP2 medium with plasmid pIJ86::sabS.

In addition to differences in pigment production by sabS and sabA mutants, we also observed morphological differences in sabS mutants when the organism was cultured on ISP2 agar medium, a rich medium in comparison to mannitol-based SGM or oatmeal-based medium. Mutants grown on ISP2 medium largely failed to produce aerial hyphae from substrate mycelium, even after prolonged growth (Fig. ​(Fig.4D).4D). Partial restoration of normal morphological development was observed in sabS mutants on ISP2 when they carried wild-type sabS on the multicopy plasmid pIJ86 (Fig. ​(Fig.4D).4D). This effect was not observed when the sabS mutant was cultured on mannitol- or starch-based medium, such as SGM (Fig. ​(Fig.4B).4B). While aerial hyphal development of sabS mutants was not impaired on SGM, we did observe a more compact, tighter colony morphology of sabS mutants on SGM plates. Partial restoration of normal colony morphology was observed in sabS mutants on SGM when they carried wild-type sabS on the multicopy plasmid pIJ86 (Fig. ​(Fig.4B4B).

The conditional morphological defect observed on ISP2 medium is reminiscent of effects seen in, e.g., S. coelicolor bldA mutants or other strains (e.g., adpA mutants) with defects in genes containing UUA codons when cultured on rich, high-osmolarity medium such as R2YE (24, 27, 33). Morphological defects can be suppressed in such mutants when cells are cultured on mannitol-based medium. The bldA gene encodes tRNA^leu and recognizes rare UUA codons in high-GC content Streptomyces mRNA transcripts. Numerous genes containing TTA codons have been identified, and many of these are genes involved in morphogenesis or secondary metabolism (5). Notably, the sabS gene sequence possesses a TTA leucine codon in the 5′ end of the gene. The codon sequence occurs at either nucleotides 13 to 15 (the 5th triplet codon) or at nucleotides 52 to 54 (the 18th triplet codon), depending on which translation start codon is used (see also Fig. ​Fig.6A).6A). Based on alignments with other GBL receptor homologs, the amino acid position occupied by leucine in SabS resides in the N terminus of the protein in a region preceding the DNA binding HTH domain. Further, this amino acid position is usually occupied by glutamic acid (12 out of 18 proteins examined); there are no other occurrences of leucine at this position in accessible GBL receptor sequences we examined. Although specific target genes subject to regulation by SabS have not yet been identified, these results suggest that SabS could function as a transcriptional regulator of genes related to morphological development. These studies are currently under way.

To more thoroughly investigate the nature of the diffusible yellow pigment compound produced in excess in sabS and sabA mutants, the mutants and wild-type strains were grown in liquid SGM, and culture filtrates were extracted using solid-phase methods. Extracts were analyzed using silica gel TLC and compared with preparations of pure WS5995B. Extracts of the ΔsabS and ΔsabA culture filtrates contained higher levels of a yellow pigment that was also present in the wild-type culture filtrate extracts, indicating overproduction of a metabolite in the mutant strains. The compound also displayed chromatographic properties on TLC plates similar to pure WS5995B. The yellow compound was recovered from preparative silica gel medium, and absorbance scans revealed that the material exhibited spectroscopic properties identical to those of pure WS5995B (data not shown). The properties of silica gel-purified compound as determined by analytical high-performance liquid chromatography were also identical to those of pure WS5995B (data not shown). While higher levels of WS5995B were found in both ΔsabS and ΔsabA mutants, the amounts of WS5995B produced by the ΔsabS mutant were substantially higher than in the ΔsabA mutant.

In order to investigate how the sabS mutation affected WS5995B production rates, growth of the wild-type and ΔsabS strains was measured over the course of WS5995B production. Detailed metabolite kinetics studies were not done with the sabA mutant. For studies of metabolite yields relative to growth in sabS mutants, it was necessary to follow growth kinetics by cell mass determinations rather than by using absorbance spectroscopy since growth of the organisms in SGM is not dispersed sufficiently enough to allow accurate analysis using spectroscopic methods. Growth kinetics of the wild-type and GBL receptor mutant ΔsabS were similar (Fig. ​(Fig.5).5). Trace amounts of WS5995B were detectable in extracts from both strains within 24 h. While the onset of WS5995B production began at approximately the same time for both strains, notable differences in the patterns and rates of production of WS5995B were seen between the wild-type and mutant strains. Production rates of WS5995B by ΔsabS varied over a 72-h time course, whereas production by the wild-type strain remained constant. After 36 h, a rapid increase in the rate of WS5995B production was observed in the mutant relative to the wild-type strain (1.36 μM·h⁻¹ for ΔsabS versus 0.1 μM·h⁻¹ for the wild type). At 48 h, the rate of production in ΔsabS decreased to that of the wild-type strain. At 60 h, WS5995B production increased again to 6 μM·h⁻¹. After about 72 h, WS5995B production in both strains reached plateau levels and began to decline soon thereafter.

Open in a separate window

FIG. 5.

Growth and WS5995B production characteristics of S. acidiscabies 84.104 wild type and a ΔsabS deletion mutant. Cell weight and WS5995B production of the wild type and the ΔsabS mutant are given, as indicated on the figure. Given large differences in WS5995B production for the wild-type and mutant strains, note that the right-hand y axis is segmented to allow inspection of WS5995B production rates for both strains. Over the 12-h interval from 36 h to 48 h, the rate of WS5995B production was 1.36 μM·h⁻¹.

SabR binds an ARE sequence element which overlaps the sabS translation initiation site.

In order to explore the biochemical properties of the GBL receptors SabR and SabS, the respective ORFs were amplified separately and cloned into the plasmid expression vector pET26b. The proteins were expressed and purified as recombinant derivatives carrying C-terminal hexahistidine sequences. Following the preparation of soluble extracts from IPTG-induced E. coli cultures, the proteins were purified using standard immobilized metal affinity chromatography methods. Both proteins were recovered from columns in elution buffer containing 300 mM imidazole. Fractions were analyzed using sodium dodecyl sulfate-PAGE, and appropriate fractions were pooled and used for mobility shift assays.

We were interested in identifying ARE DNA sequences within the sabRAS region which were recognized by SabR and/or SabS. Using pattern search tools (38), we identified two sites within or in the proximity of the sabRA intergenic region with weak but significant similarity to previously characterized ARE sequences in other streptomycetes. These sites were found centered at −12 with respect to the sabR translation start and at +14 with respect to the sabA translation start (data not shown). A radiolabeled 294-bp PCR product was generated which encompassed the sabRA intergenic region and 5′ ends of both genes, including predicted ARE sites. This labeled fragment was used in gel mobility shift assays to detect binding by SabR and/or SabS. Numerous reaction conditions were tested, and no specific binding activity was observed for either SabR or SabS, indicating that no ARE sequences are present in the sabRA intergenic region (not shown).

Pattern searching was also carried out to identify potential ARE sequences in regions upstream of sabS. Results of this analysis revealed one sequence with very strong similarity to previously characterized streptomycete ARE sequences, located in the upstream region of sabS translation initiation site (Fig. ​(Fig.6).6). A 209-bp radiolabeled PCR amplification product encompassing this ARE was used for gel mobility shift analysis to detect DNA binding activity with SabS and SabR. No DNA binding activity was detected with SabS in mobility shift assays under various conditions using this DNA substrate. The appearance of a single shifted complex was detected with SabR, however; results of these mobility shift experiments are shown in Fig. ​Fig.7.7. A shifted complex using the DNA fragment encompassing the ARE sequence within the sabS region was observed at SabR concentrations as low as 2 nM. In order to verify specific SabR binding to the ARE sequence within the 209-bp radiolabeled fragment, mobility shift assays were done with the inclusion of unlabeled duplex ARE competitor DNA. DNA-SabR complex dissociation occurred with added competitor (Fig. ​(Fig.7,7, lane 8). An apparent dissociation constant, K_D, of 14 nM was calculated from fractional binding plots of imaged mobility shift data. These data demonstrate that SabR specifically binds an AT-rich ARE box which covers the upstream region of sabS.

Open in a separate window

FIG. 7.

SabR-ARE DNA mobility shift assay. A ³²P-labeled 209-bp DNA fragment encompassing the ARE box was incubated with various amounts of recombinant SabR, and binding reaction products were electrophoresed through a 5% native PAGE gel in 0.5× Tris-borate-EDTA buffer. Lane 1, no SabR; lane 2, 1.5 nM SabR; lane 3, 3 nM SabR; lane 4, 6 nM SabR; lane 5, 12 nM SabR; lane 6, 18 nM SabR; lane 7, 24 nM SabR; lane 8, 24 nM SabR plus 150 nM unlabeled 34-bp ARE duplex competitor DNA.

Given that SabR interacts with an ARE sequence which encompasses one of two possible sabS translation initiation codons, we propose that sabS translation initiation most likely occurs at the second initiation codon, with the ARE centered at −38 with respect to translation initiation. Other lines of evidence suggest sabS is translated from this second initiation codon, resulting in production of the smaller SabS protein. Namely, translation from the 5′-most initiation codon produces an N-terminal sequence not seen in other GBL receptors, based on protein alignments (data not shown). Also, as SabR represses sabS expression (see below), the transcription level control of sabS expression could most easily be explained by binding of SabR to regions upstream of sabS translation initiation. This region also includes sequences with high probabilities of promoter elements (data not shown). These areas of inquiry are currently being investigated.

We thank the following people: Takuya Nihira and Shigeru Kitani for generously providing VB-C6 and IM-2-C6 compounds; Alaina Edmunds for constructing plasmid pALE21; Merv Bibb, Joanne Willey, and Ron Parry for providing strains and plasmids; and Justin Nodwell and Jurgen Rohr for thoughtful discussions. We also thank the anonymous reviewers of the manuscript for valuable and critical comments.

This work was supported through funding from NSF (MCB 0442509), HHMI, and Merck-AAAS.