SsaA, a Member of a Novel Class of Transcriptional Regulators,Controls Sansanmycin Production in Streptomyces sp. Strain SSthrough a Feedback Mechanism

Qinglian Li, Lifei Wang, Yunying Xie, Songmei Wang, Ruxian Chen, Bin Hong

Key Laboratory of Biotechnology of Antibiotics of Ministry of Health, Institute of Medicinal Biotechnology, Chinese Academy of Medical Sciences and Peking UnionMedical College, Beijing, China

Sansanmycins, produced by Streptomyces sp. strain SS, are uridyl peptide antibiotics with activities against Pseudomonas aerugi-nosa and multidrug-resistant Mycobacterium tuberculosis. In this work, the biosynthetic gene cluster of sansanmycins, com-prised of 25 open reading frames (ORFs) showing considerable amino acid sequence identity to those of the pacidamycin andnapsamycin gene cluster, was identified. SsaA, the archetype of a novel class of transcriptional regulators, was characterized inthe sansanmycin gene cluster, with an N-terminal fork head-associated (FHA) domain and a C-terminal LuxR-type helix-turn-helix (HTH) motif. The disruption of ssaA abolished sansanmycin production, as well as the expression of the structural genesfor sansanmycin biosynthesis, indicating that SsaA is a pivotal activator for sansanmycin biosynthesis. SsaA was proved to di-rectly bind several putative promoter regions of biosynthetic genes, and comparison of sequences of the binding sites allowedthe identification of a consensus SsaA binding sequence, GTMCTGACAN2TGTCAGKAC. The DNA binding activity of SsaA wasinhibited by sansanmycins A and H in a concentration-dependent manner. Furthermore, sansanmycins A and H were found todirectly interact with SsaA. These results indicated that SsaA strictly controls the production of sansanmycins at the transcrip-tional level in a feedback regulatory mechanism by sensing the accumulation of the end products. As the first characterized regu-lator of uridyl peptide antibiotic biosynthesis, the understanding of this autoregulatory process involved in sansanmycin biosyn-thesis will likely provide an effective strategy for rational improvements in the yields of these uridyl peptide antibiotics.

Sansanmycins, isolated from Streptomyces sp. strain SS (1, 2),are members of a uridyl peptide antibiotic family that includes

closely related pacidamycins (3), napsamycins (4), and mureido-mycins (5). These antibiotics share a structure with a 3=-deoxy-uridine nucleoside attached to an N�-methyl 2,3-diaminobutyricacid (DABA) residue of a pseudotetra/pentapeptide backbone viaan exocyclic enamide. In addition to DABA, the pseudopeptidesare comprised of other nonproteinogenic amino acids, such asmeta-Tyr (m-Tyr), and derivatives of tetradehydro-3-isoquino-line carboxylic acid (TIC) and proteinogenic amino acids Gly, Ala,Met, Trp, and Phe. In particular, the tetrapeptides of some sansan-mycins incorporate Leu to the �-amino group of DABA, which isabsent from other known uridyl peptide antibiotics (Fig. 1).Sansanmycins were reported to exert antibacterial activity againstthe highly refractory pathogen Pseudomonas aeruginosa and, moreinterestingly, Mycobacterium tuberculosis H37Rv and multidrug-resistant M. tuberculosis strains (2). As the increasing emergence ofmultidrug-resistant tuberculosis has become one of the biggestchallenges to human health worldwide, there is an urgent need todevelop novel anti-infective drugs with no cross-resistance to cur-rent clinically used antibiotics. Sansanmycins and other uridylpeptide antibiotics are of interest, as they are inhibitors of trans-locase MraY (also known as translocase I), the critical enzymeinvolved in bacterial cell wall biosynthesis, which is a clinicallyunexploited target of antibiotics (6). Elucidation of the biosyn-thetic and regulatory mechanism for these uridyl peptide antibi-otics in their producing strains will therefore facilitate the en-hancement of the production level and the generation of newbioactive uridyl peptide derivatives by genetic engineering andcombinatorial biosynthesis.

Recently, the biosynthetic gene clusters for pacidamycins (7, 8)

and napsamycins (9) were identified and characterized, suggestinga general biosynthetic pathway to the uridyl peptide antibiotics.Based on the bioinformatic analysis and in vivo and in vitro exper-iments, the assembly of pacidamycin is catalyzed by nonribosomalpeptide synthetases (NRPSs) with highly dissociated modules (7,10). The clusters also contain genes responsible for the biosynthe-sis of the DABA nonproteinogenic amino acid, the modificationof uridine nucleoside, the export of the antibiotic, and other tai-loring reactions (7–9). Generally, at least one transcriptional reg-ulatory gene lies within an antibiotic biosynthetic gene cluster,controlling biosynthesis of the respective antibiotic. However,early bioinformatic analyses did not propose a candidate regulatorfor pacidamycin biosynthesis (7, 8). In the napsamycin gene clus-ter, npsA was deduced to encode a putative regulator homologousto ArsR-type transcriptional repressors (9), but no biological ex-periments were conducted to verify its involvement in the produc-tion of napsamycins.

Streptomycetes are well known for their ability to produce var-ious secondary metabolites, including clinically used antibiotics,antitumor agents, immunosuppressants, etc. The control of sec-

Received 14 January 2013 Accepted 1 March 2013

Published ahead of print 8 March 2013

Address correspondence to Bin Hong, binhong69@hotmail.com.

Q.L. and L.W. contributed equally to this study.

Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.00054-13.

Copyright © 2013, American Society for Microbiology. All Rights Reserved.

doi:10.1128/JB.00054-13

2232 jb.asm.org Journal of Bacteriology p. 2232–2243 May 2013 Volume 195 Number 10

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

ondary metabolite production is quite complicated, involvingmultiple levels of intertwined regulation in response to physiolog-ical and environmental condition alterations (11, 12). Typically,the ultimate regulator of antibiotic production is a pathway-spe-cific transcriptional regulator situated in the respective biosyn-thetic cluster, controlling the transcription level of the corre-sponding biosynthetic genes. Most pathway-specific regulatorsbelong to the Streptomyces antibiotic regulatory protein (SARP)family, containing a DNA binding domain (DBD) close to the Nterminus and a bacterial transcriptional activation domain(BTAD) (13). The well-studied regulators, e.g., ActII-ORF4,CdaR, and RedD in S. coelicolor, controlling actinorhodin, calcium-dependent-antibiotic (CDA), and undecylprodigiosin produc-tion, respectively, belong to this family. Streptomyces LAL (largeATP binding members of the LuxR family) regulators are alsofrequent in some antibiotic gene clusters, including PikD from thepikromycin cluster of S. venezuelae (14) and multiple LAL homo-logues from the nystatin cluster of S. noursei (15). Other classes oftranscriptional regulators have also been reported recently. PimMin S. natalensis, which is highly conserved among all known poly-ene biosynthesis gene clusters (16), contains an N-terminal PASsensor domain and a C-terminal LuxR-type helix-turn-helix(HTH) domain for DNA binding (17). Atypical response regula-tors (ARRs), with an N-terminal REC domain and a winged HTHmotif in the C-terminal domain, are involved in the pathway-specific regulation of secondary metabolite production (18). The

diversity of the pathway-specific regulators may reflect the varietyof regulatory mechanism in the biosynthesis of different second-ary metabolites. JadR1, an ARR, was recently characterized to con-trol jadomycin biosynthetic gene expression in an autoregulatoryfeedback mechanism by binding of the end product, which is thefirst example of the autoregulation of antibiotic biosynthesis by anend product(s) (19). RedZ, for undecylprodigiosin (19), andAur1P, for auricin (20), were found to be modulated in a similarway, implying that end product-mediated control of antibioticpathway-specific ARRs is widespread. However, whether otherclasses of pathway-specific regulators adopt a similar mechanismremains to be elucidated.

In this work, we report the biosynthetic gene cluster of uridylpeptide antibiotic sansanmycins. The structure homology searchof SsaA, the orthologue of PacA and NpsM, revealed that it con-tained an HTH DNA binding motif at its C terminus. We furtherprovide genetic and biochemical evidence confirming that SsaA isa novel class of pathway-specific transcriptional activators ofsansanmycin biosynthesis, and that the end products can bind toand change the activity of SsaA to autoregulate antibiotic produc-tion.

MATERIALS AND METHODSStrains, media, and growth conditions. Streptomyces sp. SS (wild-typestrain) and its derivatives were grown at 28°C on S5 agar (21) for sporu-lation, on mannitol soya flour (MS) agar (22) for conjugation, in liquidfermentation medium (1) with addition of 0.1% tyrosine for sansanmycinproduction, and in liquid phage medium (23) for isolation of genomicDNA. Escherichia coli DH5� (24) was used as the host for cloning pur-poses. E. coli ET12567(pUZ8002) (25) was used to transfer DNA intoStreptomyces from E. coli by conjugation. E. coli XL1-Blue MR (Stratagene,CA) was used as the host strain for the construction of the Streptomyces sp.SS genomic DNA library. E. coli BW25113/pIJ790 was used as the host forRed/ET-mediated recombination (26). E. coli BL21(DE3) (Novagen,Madison, WI) was used as the host strain to express SsaA protein. Theywere grown at 37°C in Luria-Bertani medium (LB). For protein expres-sion in E. coli, the strain was grown in liquid ZYM-5052 medium (27). P.aeruginosa 11, the test organism for the antibacterial bioassay of sansan-mycins (2), was grown on F403 agar (21). When antibiotic selection ofbacteria was needed, strains were incubated with 50 �g ml�1 apramycin(Am), 100 �g ml�1 ampicillin (Ap), 50 �g ml�1 kanamycin (Km), 25 �gml�1 chloramphenicol (Cm), and 50 �g ml�1 streptomycin (Strep).Strains used and/or constructed in this study are listed in Table 1.

DNA manipulation. Routine DNA manipulations with E. coli werecarried out as described by Sambrook and Russell (24). RecombinantDNA techniques in Streptomyces species were performed as described byKieser et al. (22). The plasmids used and constructed in this study arelisted in Table 1. The primers for PCR amplification are shown in Table S1in the supplemental material.

Construction and screening of the genomic library. ChromosomalDNA isolated from Streptomyces sp. SS was partially digested with Sau3AIto give 40- to 50-kb DNA fragments. These fragments were dephosphor-ylated by calf intestinal alkaline phosphatase (TaKaRa, Tokyo, Japan) andthen ligated into BamHI- and HpaI-digested pOJ446. The ligation prod-ucts were packaged with Gigapack III XL Gold packaging extract (Strat-agene) by following the manufacturer’s instructions, and the resultingrecombinant phages were used to transfect E. coli XL1-Blue MR. Approx-imately 6,000 colonies were screened by colony hybridization with alka-line phosphatase-labeled ssaQ and ssaF probes, which were amplifiedfrom Streptomyces sp. SS genomic DNA and labeled using a Gene ImagesAlkPhos Direct Labeling and Detection System kit (GE Healthcare, Swe-den) according to the manufacturer’s instructions. Cosmid 13R-1 was

FIG 1 Structures of sansanmycins and some pacidamycins and napsamycins.m-Tyr, meta-Tyr; TIC, tetradehydro-3-isoquinoline carboxylic acid; MetSO,methionine sulfoxide.

SsaA Autoregulates Sansanmycin Biosynthesis

May 2013 Volume 195 Number 10 jb.asm.org 2233

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

isolated from a positive colony for subsequent complete sequencing usingthe shotgun method as performed by Majorbio (Shanghai, China).

Construction and complementation of Streptomyces sp. SS ssaAmutant. Gene disruptions were carried out by a PCR targeting method(26). Detailed methods for construction of the Streptomyces sp. SS ssaAmutant are provided in the supplemental material. The resulting mutantstrain was designated SS/AKO. A 747-bp DNA fragment containing thecomplete ssaA coding region was amplified using ssaA-pET16b-F andssaA-pICL-R (see Table S1 in the supplemental material) as primers andthen cloned into the NdeI and BamHI sites of pL646 (28) under the con-trol of a strong constitutive promoter, ermE*p. The resulting expressionvector, pL-ssaA, was introduced into SS/AKO by conjugation to give thecomplemented strain.

Analysis of sansanmycin production. Fermentation, isolation, andhigh-pressure liquid chromatography (HPLC) analysis of sansanmycinswere performed mostly by following Xie et al. (1). Detailed methods areprovided in the supplemental material.

Gene expression analysis by quantitative RT-PCR. Total RNAs wereisolated from different strains after incubation in fermentation mediumfor 48 h using a PureYield RNA Midiprep system (Promega) according tothe manufacturer’s instructions. Samples were treated with DNase I (Pro-mega) to eliminate the chromosomal DNA. The quantity and quality ofRNA samples were assessed by measuring the A260 and A280 of the samples(22). The first-strand cDNA synthesis was performed with a Turbo Scriptfirst-strand cDNA synthesis kit (Yuanpinghao, Beijng, China) using ran-dom primers by following the manufacturer’s instructions. Oligonucleo-tides were designed to amplify fragments of about 80 to 150 bp, and theprimers used are listed in Table S1 in the supplemental material. Quanti-tative real-time PCR (RT-PCR) was performed as described previously(21). Three PCR replicates were performed for each transcript, and therelative mRNA level of target genes was normalized to the level of hrdBaccording to Pfaffl’s method (29).

Expression and purification of His10-tagged SsaA. Expression andpurification of His10-tagged SsaA were performed mostly by followingWang et al. (21), with detailed protocols provided in the supplementalmaterial. Protein purity was determined by Coomassie brilliant bluestaining after SDS-PAGE on a 12% polyacrylamide gel. The concentrationof the purified SsaA was determined by the bicinchoninic acid (BCA)assay (Applygen, Beijing, China) using bovine serum albumin (BSA) as astandard.

EMSAs. Electrophoretic mobility shift assays (EMSAs) were per-formed mostly by following Wang et al. (21), with detailed protocolsprovided in the supplemental material.

DNase I footprinting. DNA fragments used in DNase I footprintingwere generated by PCR using the primers listed in Table S1 in the supple-mental material. All of these DNA fragments were cloned into pGEM-T(Promega) and then authenticated by sequencing. Probes were generatedby PCR with the universal T7 and SP6 promoter primers, with one ofthem being labeled with 5=-Alexa-647 (Invitrogen, CA). The PCR prod-ucts were purified by agarose gel electrophoresis and quantified with aNanoDrop 2000 spectrophotometer (Thermo Scientific). His10-SsaA(2.67 �M) was incubated with labeled DNA fragment (50 ng) in a totalvolume of 90 �l at 25°C for 20 min. Following several optimization ex-periments, DNase I digestions were performed with 0.1 U DNase I (gradeI; Roche, Mannheim, Germany) at 30°C for 60 s. The reactions werestopped by adding 10 �l of 500 mM EDTA (pH 8.0) and then heating at75°C for 10 min. After phenol-chloroform extraction and ethanol precip-itation, digested DNA was redissolved in 20 �l of sample loading solution(containing 0.3 �l of DNA size standard marker 600; Beckman Coulter,Brea, CA) and loaded in a GenomeLab GeXP genetic analysis system(Beckman Coulter). The results were analyzed with the GenomeLab eX-press Profiler program (Beckman Coulter).

SPR analysis. For SsaA-sansanmycin binding experiments, surfaceplasmon resonance (SPR) measurements were performed on a BIAcore

TABLE 1 Strains or plasmids used in this study

Strain or plasmid Description Reference or source

StrainsStreptomyces sp. SS Wild-type strain (sansanmycin-producing strain) 1

SS/AKO Streptomyces sp. SS with disruption of ssaA This studySS/AKO/pL-ssaA SS/AKO with the expression vector pL-ssaA; Amr This studySS/pL-ssaA Streptomyces sp. SS with the expression vector pL-ssaA; Amr This study

Escherichia coliDH5� General cloning host 24ET12567/pUZ8002 Strain used for E. coli/Streptomyces conjugation 25XL1-Blue MR Host strain for genomic library construction StratageneBW25113/pIJ790 Strain used for Red/ET-mediated recombination 26BL21(DE3) Strain used for the expression of SsaA protein Novagen

Pseudomonas aeruginosa 11 Strain used for sansanmycin bioassays 2

PlasmidspOJ446 Cosmid vector used for the construction of genomic library 4613R-1 Cosmid selected for complete sequencing via shotgun approach This study10R-1 Cosmid introduced into E. coli BW25113/pIJ790 for disruption of the minimal replicon of

SCP2* and then ssaAThis study

10R-1-SCP2KO Cosmid 10R-1 with the minimal replicon of SCP2* replaced by ampicillin resistancemarker bla; Apr, Amr

This study

10R-1-SCP2KO-AKO 10R-1-SCP2KO with ssaA replaced by streptomycin resistance gene addA; Apr, Amr, Strepr This studypIJ779 Vector used as the template for amplifying aadA cassette 26pSET152 Streptomyces integrative vector; Amr 46pL646 pSET152 derivative containing ermE*p; Amr 28pL-ssaA pL646 derivative plasmid containing 747-bp complete coding region of ssaA This studypET-16b E. coli expression vector; Apr NovagenpET-16b-A pET-16b containing the coding region of ssaA This study

Li et al.

2234 jb.asm.org Journal of Bacteriology

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

T100 system (GE Healthcare) in phosphate-buffered saline (PBS) buffer(pH 7.4, 0.05% [vol/vol] surfactant P20) with a flow rate of 30 �l min�1 at25°C. The His10-tagged SsaA was diluted to 50 �g ml�1 in 10 mM sodiumacetate (pH 5.5) and immobilized on a CM5 sensor chip by an aminocoupling kit (GE Healthcare) at a density of 8,500 response units (RU).Sansanmycin A (SS-A) and SS-H were diluted in the running buffer andinjected over the sensor chip for 3 min. Regeneration of the surface wasperformed with 5 mM NaOH for 28 s for SS-A and 0.5 mM NaOH for 16s for SS-H, followed by washes for 5 min. The data fitting for the bindingmodel was analyzed using BiaEvaluation software (GE Healthcare).

In the case of SsaA-promoter binding experiments, SPR measure-ments were performed in HBS buffer (10 mM HEPES, 150 mM NaCl, 3mM EDTA, 0.05% surfactant P20, pH 7.4) with a flow rate of 50 �l min�1

at 25°C. The biotinylated ssaC-Dp-3 fragment was diluted to 4 nM andimmobilized on an SA sensor surface at a density of 200 RU. A series ofsamples containing a fixed amount of SsaA in the presence of increasingconcentrations of SS-A was injected over the surface for 150 s. Regenera-tion of the SA sensor surface was performed with 2 M NaCl for 120 s,followed by washes for 180 s.

Bioinformatic analysis. Putative protein-coding sequences were pre-dicted using Glimmer, version 3.0 (30), and gene functional annotationwas based on BLASTP results determined with the KEGG, COG, Swiss-Port, NT, and NR databases. The program HHpred was used for struc-tural homology searches for unknown proteins (homology detection andstructure prediction by HMM-HMM comparison; http://toolkit.tuebingen.mpg.de/hhpred/). The program WebLogo was used for theidentification of the consensus binding site (http://weblogo.berkeley.edu/logo.cgi). DNA-pattern of RSA Tools (31) (http://rsat.ulb.ac.be/rsat/)was used to search a pattern (string description) within a DNA sequence.

Nucleotide sequence accession number. The sequence data of the ssacluster have been submitted to the GenBank databases under accessionnumber KC188778.

RESULTSIdentification and analysis of the sansanmycin gene cluster(ssa). Sansanmycins share a structural scaffold with pacidamycinsand napsamycins (Fig. 1); therefore, we initially attempted to lo-cate the sansanmycin gene cluster from the producer, Streptomy-ces sp. SS, through genome mining by using pacidamycin biosyn-thetic genes as probes. Nucleotide sequencing of the Streptomycessp. SS genome was undertaken by using an Illumina/Hiseq 2000sequencer (32). The size of the Streptomyces sp. SS draft genomewas predicted to be approximately 8.1 Mb, distributed over69 scaffolds containing 87 contigs (GenBank accession no.

AKXV00000000). For the local BLASTP, each open reading frame(ORF) of the pacidamycin gene cluster was used as a probe toquery against the database consisting of all of the scaffolds. Thebioinformatic search identified 14 ORFs (designated SsaH-N,SsaP-T, and SsaX-Y) on scaffold 7 and 10 ORFs (designatedSsaA-G, SsaO, and SsaV) on scaffold 36 that are putatively respon-sible for sansanmycin biosynthesis (see Fig. S1 in the supplementalmaterial). In order to obtain the complete ssa cluster, we subse-quently constructed a cosmid library of genomic DNA from Strep-tomyces sp. SS. The alkaline phosphatase-labeled fragments ofcoding sequence of ssaQ on scaffold 7 and that of ssaF on scaffold36 were used as probes to screen the library. Of �6,000 cosmids, 5gave positive signals (see Fig. S1). Three cosmids were positive forboth probes, and cosmid 13R-1 was selected for complete se-quencing via the shotgun approach. Assembly of the sequences ofscaffold 7, scaffold 36, and cosmid 13R-1 resulted in an approxi-mately 440-kb contiguous sequence in which a 33.1-kb segmentcarrying 25 ORFs is assigned to the ssa cluster (Fig. 2). ssaH andssaW are predicted to be the left and right boundaries of the ssacluster, respectively, based on the BLASTP analysis. The deducedfunction of the ORFs within the ssa cluster and five ORFs up-stream or downstream of the ssa cluster is presented in Table S2.Interestingly, the upstream and coding sequence of ssaW is almostidentical to that of ssaU (99% identity between bp 23677 and25320 and bp 36823 and 38465 in the ssa cluster).

As determined by BLASTP analysis, the ORFs of the ssa clustershow high amino acid sequence similarity to those of the pacida-mycin gene cluster (7, 8) and napsamycin gene cluster (9) and agene cluster found in S. roseosporus NRRL 15998, a known dapto-mycin producer (7). The genetic organization of the ssa clusterresembles that of the napsamycin gene cluster and the similarcluster in S. roseosporus NRRL 15998, while in the pacidamycingene cluster, pacA-G is located upstream of pacI-U. The functionsof most of the genes of the pacidamycin gene cluster have beenwell studied, especially those responsible for the peptide scaffoldassembly, and a biosynthetic pathway has been proposed based onin vivo and in vitro experiments (7, 10). All of the defined pacida-mycin biosynthetic genes, including 22 uninterrupted genes(pacA-V) and a digene cassette (pacW and pacX) identified else-where, have a homologue in the proposed ssa cluster, with only

FIG 2 Genetic organization of the ssa cluster. Genetic organization of 25 ORFs of the ssa cluster plus 5 ORFs each upstream and downstream is shown. Black linesabove the ORFs are DNA fragments retarded in the EMSA experiments, and gray lines are DNA fragments which were not retarded in the EMSA experiments.

SsaA Autoregulates Sansanmycin Biosynthesis

May 2013 Volume 195 Number 10 jb.asm.org 2235

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

one ORF (ssaY) showing homology to hypothetical genes with noknown function in the biosynthesis of a uridyl peptide compound.

SsaA is a putative pathway-specific transcriptional regulator.For the several ORFs with no information on the putative func-tions in sansanmycin biosynthesis, a structural homology searchwas performed using the online program HHpred (33). Similar toPacA (10), SsaA was predicted to contain an N-terminal forkhead-associated (FHA) domain and a C-terminal helix-turn-helix(HTH) DNA binding domain (DBD) (Fig. 3A). The FHA domainof SsaA displays the 11-stranded �-sandwich seen in other eukary-otic or prokaryotic FHA domains (Fig. 3B), which mediates theprotein-protein interaction with threonine-phosphorylated pep-tides in a sequence-specific manner (34, 35). The alignment ofFHA domains of orthologues of SsaA and the other 4 closest FHAstructural neighbors shows overall homology, but the well-knownamino acids important for the binding of phosphothreonine are

not conserved in SsaA orthologues. The DBD of SsaA displays thedomain architecture of the tetrahelical version of the HTH, with�3 as the recognition helix (Fig. 3C). The closest DBD structuralneighbors defined in HHpred were the LuxR family of transcrip-tion factors, whereas the orthologues of SsaA have longer coilsbetween the helices.

To investigate the contribution of ssaA to the regulation ofsansanmycin biosynthesis, the coding region of ssaA was clonedinto a pSET152-derived expression plasmid, pL646 (28), underthe control of a strong constitutive promoter, ermE*p, to givepL-ssaA. The plasmid was introduced into the wild-type strain byconjugation, and the resulting ssaA overexpression strain was des-ignated SS/pL-ssaA. The antibacterial bioassay against P. aerugi-nosa showed that the overexpression of ssaA increased sansanmy-cin production in SS/pL-ssaA (Fig. 4A), which was furtherconfirmed by HPLC analysis (Fig. 4B). The presence of an extra

FIG 3 Domain structure of SsaA and structure-based alignment of SsaA with its structurally homologous proteins. (A) Predicted domain structure of SsaA. (B)Structure-based alignment of the FHA domain of SsaA with its structurally homologous proteins. Three orthologues of SsaA, PacA (GenBank accession numberADN26237.1), NpsM (ADY76675.1), and SrosN15 (ZP_04694256.1), and the 4 closest FHA structural neighbors of SsaA, EmbR (PDB entry code 2FF4), yeastRad53 FHA1 (1QU5), yeast Rad53 FHA2 (1QU5), and human Ki67 (1R21), were aligned. The known conserved amino acid residues important for the bindingof phosphopeptide are shaded gray. Numbers inserted in the sequences count residues omitted from the nonconserved regions. (C) Structure-based alignmentof the DNA binding domain (DBD) of SsaA with its structurally homologous proteins. Three orthologues of SsaA, PacA (GenBank accession numberADN26237.1), NpsM (ADY76675.1), and SrosN15 (ZP_04694256.1), and the 5 closest DBD structural neighbors of SsaA, TraR (PDB entry code 1L3L), QcsR(1P4W), CviR (3QP6), RcsB (3SZT), and ComA (3ULQ), were aligned.

Li et al.

2236 jb.asm.org Journal of Bacteriology

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

copy of ssaA under ermE*p in the chromosome led to an approx-imately 100% increase in the production of sansanmycin A (SS-A)and sansanmycin H (SS-H) compared to that in the wild-typestrain containing pSET152. The results suggested that ssaA is apositive regulator of sansanmycin biosynthesis.

Disruption of ssaA completely abolished sansanmycin pro-duction. To confirm the role of ssaA in sansanmycin biosynthesis,a part of the coding region of ssaA (547 bp) was replaced with astreptomycin-resistant gene, aadA, to create the ssaA-disruptedstrain SS/AKO (see Fig. S2A in the supplemental material) accord-ing to standard methods (26). The disruption of the intendedtarget was confirmed by PCR with primers priming about 100 bpoutside the region affected by homologous recombination (seeFig. S2B) and further confirmed by Southern blotting using thestreptomycin-resistant gene aadA as the probe (see Fig. S2C). An-tibacterial activity results (Fig. 5A) and HPLC analysis (Fig. 5B)showed that disruption of ssaA totally blocked sansanmycin pro-duction. When pL-ssaA was introduced into the SS/AKO mutant,the complemented transconjugant was found to resume sansan-mycin production (Fig. 5A) at an early stage and at a higher levelthan that of the wild-type strain, which might be ascribed to theuse of the constitutive and strong promoter. These results dem-onstrated that the gene cluster identified is indeed responsible for

FIG 4 Effects of overexpression of ssaA on sansanmycin production in Strep-tomyces sp. SS. (A) Antibacterial activities of SS/pL-ssaA and SS/pSET152 atthe indicated time points. (B) HPLC analysis of sansanmycins produced bySS/pL-ssaA and SS/pSET152 at 144 h. mAU, milliabsorbance units.

FIG 5 Disruption of ssaA and its effects on sansanmycin production and biosynthetic gene expression. (A) Antibacterial activities of the wild-type strain (1),SS/AKO (2), and SS/AKO/pL-ssaA (3) at the indicated time points. (B) HPLC analysis of sansanmycins produced by the wild-type strain, SS/AKO, andSS/AKO/pL-ssaA at 144 h. (C) Transcriptional analysis of different genes in the wide-type strain, SS/AKO mutant, and its complemented strain, SS/AKO/pL-ssaA. The relative abundance at 48 h of ssaH, ssaN, ssaP, ssaA, ssaX, and ssaC transcripts in mycelia of the wide-type, SS/AKO, and SS/AKO/pL-ssaA strains wasdetermined by quantitative RT-PCR. Data are from three biological samples with one or two PCR determinations of each. The values were normalized to that ofhrdB and were represented as means � standard deviations (SD). The amounts of each particular transcript in the wild-type strain were arbitrarily assigned avalue of 1.

SsaA Autoregulates Sansanmycin Biosynthesis

May 2013 Volume 195 Number 10 jb.asm.org 2237

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

sansanmycin biosynthesis, and that ssaA is a pivotal positive reg-ulator of sansanmycin biosynthesis.

Disruption of ssaA decreased the transcription of biosyn-thetic genes. To examine further the involvement of the ssaA genein transcriptional regulation of the ssa cluster, gene expressionanalysis was conducted using quantitative RT-PCR. The relativelevels of the transcripts of five structural genes, ssaH, ssaN, ssaP,ssaX, and ssaC, located on the possible cluster boundary or as oneof the two divergent ORFs, were analyzed together with ssaA. To-tal RNAs from the wild-type, ssaA knockout mutant SS/AKO, andcomplemented SS/AKO/pL-ssaA strains were isolated under con-ditions in which the wild-type strain commenced sansanmycinproduction at about 48 h (Fig. 5A). As expected, the transcripts ofssaA were completely undetectable in the knockout mutant, whilethey were readily detectable in the wild-type strain and the com-plemented strain. The results showed that the transcripts of thestructural genes tested were almost undetectable in the knockoutstrain. In addition, the transcripts in the complemented strainwere restored to a higher level than that in the wild-type strain(Fig. 5C), consistent with the sansanmycin production level(Fig. 5A). These results indicated that ssaA could positively regu-late the sansanmycin biosynthesis by controlling the expression ofthe structural genes of the ssa cluster.

SsaA bound to the promoter regions of the ssa cluster. To bea positive regulator of sansanmycin biosynthesis, SsaA was spec-ulated to promote sansanmycin production by binding at the pro-moter regions of biosynthetic genes and thereby activating theirtranscription. Therefore, a series of EMSAs were performed toverify whether SsaA indeed interacts with the promoter regions.SsaA was overexpressed in E. coli BL21(DE3) as a His10-taggedprotein with a predicted molecular mass of 28,380 Da, and it waspurified to a purity of �90% by nickel affinity chromatography(see Fig. S3A in the supplemental material). Twelve DNA frag-ments of intergenic regions containing possible promoters of in-

terest were chosen as probes (Fig. 2), including ssaHp, ssaKp,ssaN-Pp, ssaQp, ssaX-Yp, ssaU-A-1p, ssaU-A-2p, ssaU-A-3p,ssaBp, ssaC-Dp, ssaVp, and ssaWp (the long intergenic region be-tween the divergent ssaU and ssaA was separated into three over-lapping fragments). The EMSA results showed that the purifiedHis10-tagged SsaA bound to the divergent intergenic region frag-ments ssaN-Pp, ssaU-A-1p, ssaU-A-2p, and ssaC-Dp and up-stream region fragments of two designated boundary genes, ssaHand ssaW (Fig. 6), but it did not bind to the other six fragmentsdetected (see Fig. S3B), suggesting that most of the sansanmycinbiosynthetic genes were transcribed as an operon. The bindingswere enhanced by increasing the amount of SsaA (up to 1.75 �M).Shifting of the probes was decreased when the unlabeled specificcompetitor DNA fragments were added in excess to binding reac-tions. Notably, only a single retarded band was observed forssaHp, ssaU-A-1p, ssaU-A-2p, and ssaC-Dp probes with all con-centrations of SsaA tested, whereas a second retarded band wasobserved for the ssaN-Pp probe with higher concentrations ofSsaA (0.7, 1.4, and 1.75 �M), suggesting the presence of two bind-ing sites for SsaA in this region. To restrict the binding regions toshorter segments, several subfragments were generated by PCR forssaN-Pp, ssaC-Dp, ssaU-A-1p, and ssaU-A-2p. The results showedthat two subfragments within ssaN-Pp, subfragment 3 withinssaC-Dp and subfragment B, C, and E within ssaU-Ap, were eachshifted by SsaA (see Fig. S3C). Specific binding of SsaA to severalpromoter regions of the ssa cluster in vitro, together with the re-sults of overexpression and disruption of ssaA, indicated that SsaAcontrolled sansanmycin biosynthesis by directly binding to thepromoter regions of biosynthetic genes.

Identification of SsaA binding sites in ssa promoter regions.To identify the specific binding sites of SsaA in ssa promoter re-gions, the promoter region fragments shown to be retarded inEMSAs (ssaHp, ssaC-Dp, ssaN-Pp, and ssaU-Ap-E) were studiedby DNase I footprinting analysis. Footprinting experiments were

FIG 6 EMSA analysis of SsaA with the postulated promoter regions of the ssa cluster. Data represent EMSA analysis of 3= biotin-11-ddUTP-labeled fragmentsssaHp, ssaN-Pp, ssaC-Dp, ssaU-A-1p, ssaU-A-2p, and ssaWp with His10-SsaA. For the ssaHp, ssaN-Pp, ssaU-A-1p, ssaU-A-2p, and ssaC-Dp images, the minusindicates probe only. Lanes under the triangle indicate increasing concentrations of SsaA (0, 0.175, 0.35, 0.7, 1.4, and 1.75 �M) incubated with the probe. Cindicates 0.7 �M His10-SsaA incubated with a 200-fold excess of unlabeled specific competitor DNA fragment. In the case of ssaWp, the minus indicates probeonly, and a plus indicates probe incubated with 0.175 �M His10-SsaA.

Li et al.

2238 jb.asm.org Journal of Bacteriology

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

performed on both the coding strand and the complementarystrand of the DNA fragments (Fig. 7A; also see Fig. S4 in thesupplemental material).

Footprinting analysis of the ssaHp region revealed a 31-nucle-otide protected sequence in the coding strand (positions �135 to�105 relative to the translation start site of ssaH). In the comple-mentary strand, the protected sequence was 13 nucleotides long,at positions �118 to �106. These two protected regions over-lapped by 12 nucleotides (see Fig. S4A in the supplemental mate-rial).

Results of the analysis of the ssaC-Dp region revealed that SsaAprotected a region of 28 nucleotides, from bp �238 to �211 rel-ative to the ssaD translation start site in the coding strand. In the

complementary strand, SsaA protected a 21-bp sequence, locatedat nucleotide positions �228 to �208 relative to the ssaD transla-tion start site (see Fig. S4B in the supplemental material).

In the case of the ssaU-Ap-E fragment, the footprint on thecoding strand covered a region of 25 nucleotides, from bp �411 to�387 relative to the ssaU translation start site. A 16-nucleotidefootprint of SsaA binding at the complementary strand was ob-tained, spanning from positions �385 to �370 relative to the ssaUtranslation start site. The two protected regions did not overlap atall (see Fig. S4C in the supplemental material).

Footprinting analysis with the ssaN-Pp region revealed twoprotected areas in each strand, in accordance with the appearanceof two retarded bands in EMSA experiments. In the coding strand,

FIG 7 Identification of SsaA binding sites. (A) Identification of SsaA binding site in ssaN-Pp promoter region by DNase I footprinting. The upper electrophero-gram shows the control reaction, and the lower one shows DNase I footprints of His10-SsaA (2.67 �M) bound to the labeled DNA fragments (50 ng). Theprotected nucleotide sequences are shaded gray. The nucleotide positions of the protected sequences respective to the ssaP translation start point are shown belowthe gray shading. (B) Identification of the SsaA consensus binding site by WebLogo. Six sequences were aligned, including 5 sequences detected by the DNase Ifootprinting experiments and 1 sequence upstream of ssaW that is identical to that of ssaU. The base-pairing nucleotides within the binding site are underlined.(C) Validation of the identified SsaA consensus binding site. The binding of SsaA to the three synthetic duplexes, D1, D2, and D3 (the dashed lines in D3 indicatethe deletion of 5 nucleotides), was detected by EMSA analysis. �, probe only; �, probe incubated with 1.75 �M His10-SsaA.

SsaA Autoregulates Sansanmycin Biosynthesis

May 2013 Volume 195 Number 10 jb.asm.org 2239

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

SsaA protected two regions, stretching from positions �320 to�295 and from �193 to �173 with respect to the ssaP translationstart site. In the complementary strand, two protected regions,stretching from positions �286 to �271 and from �175 to �156relative to the ssaP translation start site, were observed (Fig. 7A).

It is interesting that the SsaA footprints on the coding strandare not accompanied by totally equivalent footprints on the com-plementary strand for all four promoter regions detected, but byaligning the five sequences within the protected regions of bothstrands and the homologous sequence in ssaWp, an SsaA consen-sus binding sequence of GTMCTGACAN2TGTCAGKAC (whereM represents A or C and K is G or T) was identified (Fig. 7B),which consisted of two 9-bp inverted repeats (IRs; indicated byunderlining) separated by a 2-bp linker. A sequence logo (36) thatdepicts the binding sites is shown in Fig. 7B.

To validate the identified SsaA consensus binding site, we de-signed three duplexes, one containing the canonical binding site

(GTCCTGACAGCTGTCAGGAC; D1), a second one in whichCTGAC of one of the IRs was replaced by ATTCA (GTCATTCAAGCTGTCAGGAC; D2), and a third one with GTCAG of one ofthe IRs deleted (GTCCTGACAGCT-----GAC; D3) (underliningindicates the nucleotides that were mutated or deleted in D2 andD3). We examined the binding of SsaA to these three duplexes inEMSA experiments. The results showed that SsaA interacts withthe D1 duplex (Fig. 7C). These results validated the identified SsaAconsensus binding site and implied that the two IRs both are im-portant for the binding activity of SsaA.

Sansanmycins modulate the DNA binding activity of SsaA byinteracting with SsaA. To test whether the sansanmycin endproducts have an effect on the DNA binding activity of SsaA,EMSAs were performed using the promoter probes ssaC-Dp andssaU-A-1p in the presence of different amounts of SS-A or SS-H(Fig. 8). The results showed that SS-A inhibited the band shiftingof ssaC-Dp and ssaU-A-1p, which was caused by SsaA in a concen-

FIG 8 Effect of sansanmycins on the DNA binding activity of SsaA. (A) Effect of different amounts of SS-A on the binding of His10-SsaA (0.175 �M) to ssaC-Dpor ssaU-A-1p. Lane 1, probe only; lane 2, probe incubated with 0.175 �M His10-SsaA; lanes 3 to 10, probe incubated with 0.175 �M His10-SsaA in the presenceof increasing concentrations of SS-A (50, 100, 200, 300, 450, 600, 800, and 1,000 �M). (B) Effect of different amounts of SS-H on the binding of His10-SsaA (0.175�M) to ssaC-Dp or ssaU-A-1p. Lane 1, probe only; lane 2, probe incubated with 0.175 �M His10-SsaA; lanes 3 to 10, probe incubated with 0.175 �M His10-SsaAin the presence of increasing concentrations of SS-H (50, 100, 200, 300, 450, 600, 800, and 1,000 �M). (C) SPR analysis of the effect of SS-A on the DNA bindingactivity of SsaA. The binding of His10-SsaA (100 nM) to ssaC-Dp-3 fragment immobilized on an SA chip was inhibited by the increasing concentrations of SS-A(50, 100, 200, and 300 �M). (D) SPR analysis of SS-A interaction with immobilized His10-SsaA. Sensorgrams for the increasing concentrations of SS-A (3.125,6.25 12.5, 25, and 50 �M) show its binding to the immobilized His10-SsaA on a CM5 chip. (E) SPR analysis of SS-H interaction with immobilized His10-SsaA.Sensorgrams for the increasing concentrations of SS-H (12.5, 25, 50, 100, and 200 �M) show its binding to the immobilized His10-SsaA on a CM5 chip.

Li et al.

2240 jb.asm.org Journal of Bacteriology

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

tration-dependent manner (Fig. 8A), while SS-H did it in a similarmanner (Fig. 8B). These results were further confirmed by SPRanalysis, as shown in Fig. 8C. A concentration-dependent inhibi-tion of SsaA binding to the ssaC-Dp-3 fragment immobilized onan SA sensor chip was readily observed when the amount of SS-Awas increased. All of these results suggested that sansanmycinscould modulate the binding activity of SsaA for its target DNA.

To determine whether sansanmycins modulate the DNA bind-ing activity of SsaA through direct interaction with SsaA, SPRexperiments were conducted using a CM5 sensor chip. Over aconcentration range of 3.125 to 200 �M SS-A, SPR sensorgramsshowed an increasing binding of SS-A to the immobilized SsaA(Fig. 8D). A similar result was obtained for the SPR analysis of theinteraction of SS-H with immobilized SsaA (Fig. 8E). The best fitsfor both SS-A and SS-H were obtained with the two-stage reactionmodel, yielding a KD (equilibrium dissociation constant) value of318.1 and 781.7 �M for SS-A and SS-H, respectively, suggestingthat the binding of a first molecule allows, by inducing a confor-mational change, the binding of a second molecule.

DISCUSSION

The sansanmycin biosynthetic gene cluster was identified in Strep-tomyces sp. SS by a combination of genome mining and conven-tional probing and cosmid sequencing. Consistent with the highsimilarity of the structures of sansanmycins to those of pacidamy-cins and napsamycins, the three clusters share a vast majority ofhomologues (ssaA-V). The ssa and nps clusters present almost thesame genetic organization, except for several unidentified smallORFs. They both contain the homologue of pacX responsible forthe synthesis of m-Tyr in the pseudopeptide, which was identifiedelsewhere in a pacWX digene cassette in S. coeruleorubidus NRRL18370 (10). However, the homologue of NpsU, responsible for thereduced uracil moiety in napsamycin (9), was not found in ssa andpac clusters. Interestingly, the ssa cluster contains an extra ORF(SsaW) at the right boundary of the cluster. SsaW is predicted toencode a freestanding A domain of NRPS, showing 74% identityto PacU and 77% identity to PacW from S. coeruleorubidus NRRL18370. PacU was demonstrated to specifically activate the N-ter-minal Ala found in pacidamycin D/S (7), and PacW was found toactivate the N-terminal m-Tyr found in pacidamycin 4/5/5T (10).However, SsaW matches almost perfectly with the amino acidsequence of SsaU (with only one alteration of Lys to Arg), suggest-ing that SsaW is the early duplicate of SsaU in the cluster and isevolutionarily viable for incorporation of a new kind of aminoacid. Unlike the nps cluster, which has an ArsR-type regulatoraround the cluster boundary, there is no homologue to any kind oftranscriptional regulator in the near upstream or downstream ofthe defined ssa cluster.

Our studies have identified a new class of transcriptional fac-tor, SsaA, which functions as a pivotal activator of the sansanmy-cin production in Streptomyces sp. SS. This was established bytargeted ssaA gene disruption and complementation together withtranscript analysis. It was confirmed when the overexpression ofSsaA in the wild-type strain and complementation of SsaA in thessaA disruption strain under a strong promoter both significantlyincreased the sansanmycin production level. This indicates thatthe endogenous SsaA is not present in saturating amounts, andelevating the expression level of ssaA could be a practical strategyto improve uridyl peptide antibiotic production. To the best ofour knowledge, SsaA is the first-described example of a regulator

that combines an N-terminal fork head-associated (FHA) domainwith a C-terminal helix-turn-helix (HTH) motif of the LuxR typefor DNA binding. SsaA bears no significant homology to any char-acterized gene products in the NCBI database according toBLASTP searches. In addition to there being orthologues in re-ported uridyl peptide antibiotic clusters with �80% identity,there are some hypothetical proteins from actinomycetes, mainlystreptomycetes, showing considerable similarity (�30% overallidentity), suggesting that this class of regulators is generally in-volved in cellular physiology, including secondary metabolite bio-synthesis in actinomycetes.

EMSAs were used here to prove the direct binding of SsaA tocertain promoters of the ssa genes. Binding of the SsaA protein wasdemonstrated at 6 out of 12 postulated intergenic fragments.More than one shift band was obtained with ssaN-Pp at increasingprotein concentrations, indicating that more than one bindingsite is present in this region, a result that was confirmed by EMSAof subfragments of this region. The lack of binding to the up-stream region of ssaKp, ssaQp, ssaBp, and ssaVp, which is in thesame direction as its upstream gene and has a relatively long in-tergenic region, suggests that most, if not all, blocks of codirec-tional genes are controlled as operons. Thus, it appears plausiblethat transcription from most of the sansanmycin biosyntheticgenes can be controlled by SsaA. This is a notable feature of somepathway-specific regulators of antibiotic biosynthesis, minimiz-ing the number of promoters by extensive usage of operon con-trol, as exemplified by control of streptomycin production by StrRin S. griseus (37).

DNase I footprinting experiments revealed protected areas ofapproximately 20 to 30 nucleotides in the target promoters. Typ-ically, the protected region in the sense strand of the activated geneis accompanied by a slightly displaced protection in the comple-mentary strand, suggesting one monomer of the protein binds oneface of the DNA helix. A comparison of the protected sequencesled to the identification of a conserved palindromic sequencespanning 20 nucleotides. The relevance of the IRs as the cis regu-latory element of SsaA was demonstrated by mutation or deletionof the most conserved CTGAC sequence in the IRs. The dyadsymmetry of the binding site of SsaA is in agreement with thebinding sites of regulators of the LuxR type, such as LuxR (38) andTraR (39). The presence of highly conserved SsaA orthologues inall of the reported uridyl peptide antibiotic biosynthetic clusterssuggests the relevance of this new class of transcriptional regula-tors in the regulation of their production. Interestingly, the otherreported uridyl peptide antibiotic clusters also contain the highlyconserved SsaA binding sites in the divergent intergenic regions orthe upstream region of genes on the cluster boundary (see TableS3 in the supplemental material), proving the general applicabilityof the consensus binding site of these transcriptional regulators.

Gene expression analysis in the wild-type strain and SS/AKOmutant by quantitative RT-PCR revealed that the transcription ofssaX was dependent on SsaA. However, EMSA results showed thatSsaA cannot bind to the intergenic region between ssaX and ssaY.Sequence search of this region and further upstream region to theend of ssaU did not find the conserved SsaA binding site either.One explanation for these results is that the transcription of ssaX iscontrolled by another hierarchical regulator, which would be ac-tivated by SsaA. Another possibility is that ssaX is cotranscribedwith ssaU, which is located in the far upstream of ssaX. All of thesehypotheses required further experimental tests.

SsaA Autoregulates Sansanmycin Biosynthesis

May 2013 Volume 195 Number 10 jb.asm.org 2241

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

In more detailed EMSA analyses, the major end products ofsansanmycins were found to markedly inhibit the DNA bindingactivity of SsaA to the two promoter regions tested. This result,further confirmed by an SPR analysis, raised the possibility thatthe end products of sansanmycins modulate SsaA activity. How-ever, sansanmycins overproduced in a pL-ssaA overexpressionstrain did not reduce the transcription of sansanmycin biosynthe-sis genes. It is possible that the elevated SsaA protein levels in thepL-ssaA-bearing strain are much higher than the concentrationrequired for the titration of the sansanmycin end products. Thedirect interaction of SS-A and SS-H with SsaA was then demon-strated by SPR analysis, indicating that the DNA binding activityof SsaA could be directly modulated by the end products ofsansanmycins. It is well known that the FHA domain is responsi-ble for the interaction of phosphoprotein containing a phospho-threonine (pThr), which is involved in diverse biological pathways(34). The typical FHA domain comprises approximately 75 aminoacids for pThr peptide recognition involving the highly conservedresidues Arg, Ser, and Asn. However, SsaA resembles some FHAdomains functioning as a non-pThr binding module (40), withthe only invariant residue, Gly, following strand �3. As each uridylpeptide antibiotic consists of a pseudopeptide, it seems reasonableto speculate that the end products can bind to the FHA domain tochange the conformation of SsaA, which influence its DNA bind-ing activity.

Low-molecular-weight compounds, including intermediatesor end products of the respective biosynthetic pathways, were pre-viously reported to influence antibiotic production and export.For example, in tylosin biosynthesis, glycosylated macrolides wereproposed to exert a pronounced positive effect on polyketide me-tabolism in S. fradiae (41). Rhodomycin D, the first glycosylatedintermediate in donorubicin biosynthetic pathways, can interactwith the TetR-like pathway-specific regulator DnrO to relieve itsself-repression, thereby positively enhancing end production(42). A feed-forward mechanism has been proposed by Tahlan etal. (43) in which an antibiotically inactive precursor relieves re-pression by the TetR-like regulator ActR, ensuring the expressionof the efflux pump ActA to couple the antibiotic biosynthesis andexport. A similar regulation of efflux of simocyclinone was re-ported by Le et al. (44). Here, we demonstrated that the novel classof FHA-LuxR-type regulator, SsaA, adopted a negative autoregu-lation strategy to strictly control the production of uridyl peptideantibiotics, regulating the biosynthetic genes at the transcriptionallevel by sensing the concentration of the end products of the bio-synthetic pathway. This strategy might ensure that the producerorganism does not inhibit its own growth. A similar autoregula-tory mechanism was recently reported in which end products andsome intermediates in jadomycin biosynthesis can bind to itspathway-specific regulator, an atypical response regulator (ARR),and affect its binding activity (19). This mechanism also hasbeen found in auricin (20) and simocyclinone (45) production.The negative autoregulation by end products may be a wide-spread regulatory mechanism for antibiotic biosynthesis instreptomycetes.

In summary, we have identified and characterized a member ofa novel class of transcriptional regulators with the FHA domain inthe global regulation of the sansanmycin biosynthesis. Knowledgeof this regulator may set the stage for an understanding of thegenetic control of uridyl peptide antibiotic biosynthesis and its

regulation and provide an effective strategy to improve the yieldsof these antibiotics.

ACKNOWLEDGMENTS

We gratefully acknowledge K. McDowall for plasmid pL646 and Entai Yaoand Ning He for their help in the fermentation of the SS strains.

This work was supported by the National Natural Science Foundationof China (31170042, 30973668, and 81273415), the National Megaprojectfor Innovative Drugs (2012ZX09301002-001-016, 2009ZX09501-008,and 2010ZX09401-403), the Beijing Natural Science Foundation(5102032), and the Fundamental Research Funds for the Central Univer-sities (2012N09).

REFERENCES1. Xie Y, Chen R, Si S, Sun C, Xu H. 2007. A new nucleosidyl-peptide

antibiotic, sansanmycin. J. Antibiot. (Tokyo) 60:158 –161.2. Xie Y, Xu H, Si S, Sun C, Chen R. 2008. Sansanmycins B and C, new

components of sansanmycins. J. Antibiot. (Tokyo) 61:237–240.3. Chen RH, Buko AM, Whittern DN, McAlpine JB. 1989. Pacidamycins,

a novel series of antibiotics with anti-Pseudomonas aeruginosa activity. II.Isolation and structural elucidation. J. Antibiot. (Tokyo) 42:512–520.

4. Chatterjee S, Nadkarni SR, Vijayakumar EK, Patel MV, Ganguli BN,Fehlhaber HW, Vertesy L. 1994. Napsamycins, new Pseudomonas activeantibiotics of the mureidomycin family from Streptomyces sp. HILY-82,11372. J. Antibiot. (Tokyo) 47:595–598.

5. Isono F, Inukai M, Takahashi S, Haneishi T, Kinoshita T, Kuwano H.1989. Mureidomycins A-D, novel peptidylnucleoside antibiotics withspheroplast forming activity. II. Structural elucidation. J. Antibiot. (To-kyo) 42:667– 673.

6. Winn M, Goss RJ, Kimura K, Bugg TD. 2010. Antimicrobial nucleosideantibiotics targeting cell wall assembly: recent advances in structure-function studies and nucleoside biosynthesis. Nat. Prod. Rep. 27:279 –304.

7. Zhang W, Ostash B, Walsh CT. 2010. Identification of the biosyntheticgene cluster for the pacidamycin group of peptidyl nucleoside antibiotics.Proc. Natl. Acad. Sci. U. S. A. 107:16828 –16833.

8. Rackham EJ, Gruschow S, Ragab AE, Dickens S, Goss RJ. 2010. Pacida-mycin biosynthesis: identification and heterologous expression of the firsturidyl peptide antibiotic gene cluster. Chembiochem 11:1700 –1709.

9. Kaysser L, Tang X, Wemakor E, Sedding K, Hennig S, Siebenberg S,Gust B. 2011. Identification of a napsamycin biosynthesis gene cluster bygenome mining. Chembiochem 12:477– 487.

10. Zhang W, Ntai I, Bolla ML, Malcolmson SJ, Kahne D, Kelleher NL,Walsh CT. 2011. Nine enzymes are required for assembly of the pacida-mycin group of peptidyl nucleoside antibiotics. J. Am. Chem. Soc. 133:5240 –5243.

11. Martin JF, Liras P. 2010. Engineering of regulatory cascades and net-works controlling antibiotic biosynthesis in Streptomyces. Curr. Opin. Mi-crobiol. 13:263–273.

12. van Wezel GP, McDowall KJ. 2011. The regulation of the secondarymetabolism of Streptomyces: new links and experimental advances. Nat.Prod. Rep. 28:1311–1333.

13. Wietzorrek A, Bibb M. 1997. A novel family of proteins that regulatesantibiotic production in streptomycetes appears to contain an OmpR-likeDNA-binding fold. Mol. Microbiol. 25:1181–1184.

14. Wilson DJ, Xue Y, Reynolds KA, Sherman DH. 2001. Characterizationand analysis of the PikD regulatory factor in the pikromycin biosyntheticpathway of Streptomyces venezuelae. J. Bacteriol. 183:3468 –3475.

15. Sekurova ON, Brautaset T, Sletta H, Borgos SE, Jakobsen MO, Elling-sen TE, Strom AR, Valla S, Zotchev SB. 2004. In vivo analysis of theregulatory genes in the nystatin biosynthetic gene cluster of Streptomycesnoursei ATCC 11455 reveals their differential control over antibiotic bio-synthesis. J. Bacteriol. 186:1345–1354.

16. Santos-Aberturas J, Payero TD, Vicente CM, Guerra SM, Canibano C,Martin JF, Aparicio JF. 2011. Functional conservation of PAS-LuxRtranscriptional regulators in polyene macrolide biosynthesis. Metab. Eng.13:756 –767.

17. Santos-Aberturas J, Vicente CM, Guerra SM, Payero TD, Martin JF,Aparicio JF. 2011. Molecular control of polyene macrolide biosynthesis:direct binding of the regulator PimM to eight promoters of pimaricingenes and identification of binding boxes. J. Biol. Chem. 286:9150 –9161.

18. Guthrie EP, Flaxman CS, White J, Hodgson DA, Bibb MJ, Chater KF.

Li et al.

2242 jb.asm.org Journal of Bacteriology

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

1998. A response-regulator-like activator of antibiotic synthesis fromStreptomyces coelicolor A3(2) with an amino-terminal domain that lacks aphosphorylation pocket. Microbiology 144(Pt 3):727–738.

19. Wang L, Tian X, Wang J, Yang H, Fan K, Xu G, Yang K, Tan H. 2009.Autoregulation of antibiotic biosynthesis by binding of the end product toan atypical response regulator. Proc. Natl. Acad. Sci. U. S. A. 106:8617–8622.

20. Kutas P, Feckova L, Rehakova A, Novakova R, Homerova D, MingyarE, Rezuchova B, Sevcikova B, Kormanec J. 2012. Strict control of auricinproduction in Streptomyces aureofaciens CCM 3239 involves a feedbackmechanism. Appl. Microbiol. Biotechnol. 97:2413–2421.

21. Wang L, Hu Y, Zhang Y, Wang S, Cui Z, Bao Y, Jiang W, Hong B. 2009.Role of sgcR3 in positive regulation of enediyne antibiotic C-1027 produc-tion of Streptomyces globisporus C-1027. BMC Microbiol. 9:14. doi:10.1186/1471-2180-9-14.

22. Kieser T, Bibb MJ, Buttner MJ, Chater KF, Hopwood DA. 2000. PracticalStreptomyces genetics. The John Innes Foundation, Norwich, CT.

23. Korn F, Weingartner B, Kutzner HJ. 1978. A study of twenty acti-nophages: morphology, serological relationship and host range, p 251–270. In Freerksen E, Tarnok I, Thumin H (ed), Genetics of the Actinomy-cetales. G. Fisher, New York, NY.

24. Sambrook J, Russell DW. 2001. Molecular cloning: a laboratory manual.Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.

25. Paget MS, Chamberlin L, Atrih A, Foster SJ, Buttner MJ. 1999. Evidencethat the extracytoplasmic function sigma factor sigmaE is required fornormal cell wall structure in Streptomyces coelicolor A3(2). J. Bacteriol.181:204 –211.

26. Gust B, Challis GL, Fowler K, Kieser T, Chater KF. 2003. PCR-targetedStreptomyces gene replacement identifies a protein domain needed forbiosynthesis of the sesquiterpene soil odor geosmin. Proc. Natl. Acad. Sci.U. S. A. 100:1541–1546.

27. Studier FW. 2005. Protein production by auto-induction in high densityshaking cultures. Protein Expr. Purif. 41:207–234.

28. Hong B, Phornphisutthimas S, Tilley E, Baumberg S, McDowall KJ.2007. Streptomycin production by Streptomyces griseus can be modulatedby a mechanism not associated with change in the adpA component of theA-factor cascade. Biotechnol. Lett. 29:57– 64.

29. Pfaffl MW. 2001. A new mathematical model for relative quantification inreal-time RT-PCR. Nucleic Acids Res. 29:e45.

30. Delcher AL, Bratke KA, Powers EC, Salzberg SL. 2007. Identifyingbacterial genes and endosymbiont DNA with Glimmer. Bioinformatics23:673– 679.

31. van Helden J. 2003. Regulatory sequence analysis tools. Nucleic AcidsRes. 31:3593–3596.

32. Wang L, Xie Y, Li Q, He N, Yao E, Xu H, Yu Y, Chen R, Hong B. 2012.Draft genome sequence of Streptomyces sp. SS, which produces a series ofuridyl peptide antibiotic sansanmycins. J. Bacteriol. 194:6988 – 6989.

33. Soding J, Biegert A, Lupas AN. 2005. The HHpred interactive server forprotein homology detection and structure prediction. Nucleic Acids Res.33:W244 –W248.

34. Liang X, van Doren SR. 2008. Mechanistic insights into phosphoprotein-binding FHA domains. Acc. Chem. Res. 41:991–999.

35. Hammet A, Pike BL, McNees CJ, Conlan LA, Tenis N, Heierhorst J.2003. FHA domains as phospho-threonine binding modules in cell signal-ing. IUBMB Life 55:23–27.

36. Crooks GE, Hon G, Chandonia JM, Brenner SE. 2004. WebLogo: asequence logo generator. Genome Res. 14:1188 –1190.

37. Retzlaff L, Distler J. 1995. The regulator of streptomycin gene expression,StrR, of Streptomyces griseus is a DNA binding activator protein with mul-tiple recognition sites. Mol. Microbiol. 18:151–162.

38. Pompeani AJ, Irgon JJ, Berger MF, Bulyk ML, Wingreen NS, BasslerBL. 2008. The Vibrio harveyi master quorum-sensing regulator, LuxR, aTetR-type protein is both an activator and a repressor: DNA recognitionand binding specificity at target promoters. Mol. Microbiol. 70:76 – 88.

39. Zhu J, Winans SC. 1999. Autoinducer binding by the quorum-sensingregulator TraR increases affinity for target promoters in vitro and de-creases TraR turnover rates in whole cells. Proc. Natl. Acad. Sci. U. S. A.96:4832– 4837.

40. Tong Y, Tempel W, Wang H, Yamada K, Shen L, Senisterra GA,MacKenzie F, Chishti AH, Park HW. 2010. Phosphorylation-independent dual-site binding of the FHA domain of KIF13 mediatesphosphoinositide transport via centaurin alpha1. Proc. Natl. Acad. Sci.U. S. A. 107:20346 –20351.

41. Fish SA, Cundliffe E. 1997. Stimulation of polyketide metabolism inStreptomyces fradiae by tylosin and its glycosylated precursors. Microbiol-ogy 143(Pt 12):3871–3876.

42. Jiang H, Hutchinson CR. 2006. Feedback regulation of doxorubicinbiosynthesis in Streptomyces peucetius. Res. Microbiol. 157:666 – 674.

43. Tahlan K, Ahn SK, Sing A, Bodnaruk TD, Willems AR, Davidson AR,Nodwell JR. 2007. Initiation of actinorhodin export in Streptomyces coe-licolor. Mol. Microbiol. 63:951–961.

44. Le TB, Fiedler HP, den Hengst CD, Ahn SK, Maxwell A, Buttner MJ.2009. Coupling of the biosynthesis and export of the DNA gyrase inhibitorsimocyclinone in Streptomyces antibioticus. Mol. Microbiol. 72:1462–1474.

45. Horbal L, Rebets Y, Rabyk M, Makitrynskyy R, Luzhetskyy A,Fedorenko V, Bechthold A. 2012. SimReg1 is a master switch for biosyn-thesis and export of simocyclinone D8 and its precursors. AMB Express2:1.

46. Bierman M, Logan R, O’Brien K, Seno ET, Rao RN, Schoner BE. 1992.Plasmid cloning vectors for the conjugal transfer of DNA from Escherichiacoli to Streptomyces spp. Gene 116:43– 49.

SsaA Autoregulates Sansanmycin Biosynthesis

May 2013 Volume 195 Number 10 jb.asm.org 2243

on February 12, 2018 by guest

http://jb.asm.org/

Dow

nloaded from