Identiﬁcation of PimR as a Positive Regulator of Pimaricin ... · rapamycin pathway in...

JOURNAL OF BACTERIOLOGY, May 2004, p. 2567–2575 Vol. 186, No. 90021-9193/04/$08.00�0 DOI: 10.1128/JB.186.9.2567–2575.2004Copyright © 2004, American Society for Microbiology. All Rights Reserved.

Identification of PimR as a Positive Regulator of PimaricinBiosynthesis in Streptomyces natalensis

Nuria Anton,1 Marta V. Mendes,1 Juan F. Martín,1,2 and Jesus F. Aparicio1,2*Institute of Biotechnology INBIOTEC, Parque Científico de Leon, 24006 Leon,1 and Area of

Microbiology, Faculty of Biology, University of Leon, 24071 Leon,2 Spain

Received 4 December 2003/Accepted 15 January 2004

Sequencing of the DNA region on the left fringe of the pimaricin gene cluster revealed the presence of a3.6-kb gene, pimR, whose deduced product (1,198 amino acid residues) was found to have amino acid sequencehomology with bacterial regulatory proteins. Database comparisons revealed that PimR represents the arche-type of a new class of regulators, combining a Streptomyces antibiotic regulatory protein (SARP)-like N-terminal section with a C-terminal half homologous to guanylate cyclases and large ATP-binding regulators ofthe LuxR family. Gene replacement of pimR from Streptomyces natalensis chromosome results in a complete lossof pimaricin production, suggesting that PimR is a positive regulator of pimaricin biosynthesis. Gene expres-sion analysis by reverse transcriptase PCR (RT-PCR) of the pimaricin gene cluster revealed that S. natalensis�PimR shows no expression at all of the cholesterol oxidase-encoding gene pimE, and very low level tran-scription of the remaining genes of the cluster except for the mutant pimR gene, thus demonstrating that thisregulator activates the transcription of all the genes belonging to the pimaricin gene cluster but not its owntranscription.

Streptomycetes are filamentous soil bacteria that have acomplex life cycle that involves differentiation and sporulation.These bacteria have attracted great interest due to their well-known ability to produce a great variety of antibiotics andother secondary metabolites. Production of these compoundsis regulated in response to nutritional status alteration and avariety of environmental conditions and, hence, occurs in agrowth-phase-dependent manner, at the beginning of the sta-tionary growth phase, and is usually accompanied by morpho-logical differentiation (10, 15).

Control of secondary metabolite production seems to beoperating at several levels. The highest levels include genesthat exert a pleiotropic control over one or more aspects ofsecondary metabolism, such as antibiotic production or mor-phological differentiation (see reference 16 for a review). Thelowest level, however, is composed by regulatory genes thatonly affect a single antibiotic biosynthetic pathway. These path-way-specific regulatory genes are usually found within the re-spective antibiotic biosynthesis gene cluster, a feature that hasgreatly facilitated their study.

The first pathway-specific regulatory proteins characterizedin actinomycetes belonged to a protein family called SARPs(Streptomyces antibiotic regulatory proteins) (43). Members ofthis family are characterized by the presence of an OmpR-likeDNA-binding domain (31) and include ActII-Orf4 of the ac-tinorhodin pathway (5), DnrI of the daunorubicin gene cluster(28), RedD of the undecylprodigiosin pathway (33), CcaR ofthe cephamycin and clavulanic acid pathways (36), and TylSand TylT from the tylosin pathway (8), among others.

In recent years, a novel family of transcriptional regulators

has been identified (20). This new family, typified by the reg-ulator of the maltose regulon in Escherichia coli, MalT (11),and called large ATP-binding regulators of the LuxR family(LAL) is characterized by the unusual size of its members, thepresence of an N-terminal ATP/GTP-binding domain easilyidentified by the presence of the conserved Walker A motif(42), and a C-terminal LuxR-like DNA-binding domain char-acterized by a conserved helix-turn-helix motif. Several regu-lators of the LAL family have been identified in antibiotic geneclusters from actinomycetes, including PikD from the pikro-mycin pathway in Streptomyces venezuelae (44), RapH from therapamycin pathway in Streptomyces hygroscopicus (1, 32), andNysRI and NysRIII from the nystatin pathway in Streptomycesnoursei (12).

Pimaricin is a 26-member tetraene macrolide antifungal an-tibiotic produced by Streptomyces natalensis, which is widelyused for the treatment of fungal keratitis and also in the foodindustry to prevent mold contamination of cheese and othernonsterile foods (i.e., cured meat, sausages, and ham, etc.). Asa polyene, its antifungal activity lies in its interaction withmembrane sterols, thus causing the alteration of membranestructure and leading to the leakage of cellular materials. Asother macrocyclic polyketides, pimaricin is synthesized by theaction of so-called type I modular polyketide synthases (4).Our laboratory has previously sequenced the pimaricin biosyn-thetic gene cluster completely and demonstrated its identity bygene disruption experiments (2, 3). The gene cluster encodes13 polyketide synthase modules within five multifunctional en-zymes and 12 additional proteins that presumably govern mod-ification of the polyketide skeleton, export, and regulation ofgene expression (Fig. 1) (3, 4).

In this paper we describe the cloning, sequencing, and de-tailed characterization of a novel class of pathway-specific reg-ulators in S. natalensis and demonstrate its role as a transcrip-tional activator for pimaricin biosynthesis in this bacterium.

* Corresponding author. Mailing address: Institute of BiotechnologyINBIOTEC, Parque Científico de Leon, Avda. del Real, no. 1, 24006Leon, Spain. Phone: 34-987-210308. Fax: 34-987-210388. E-mail:[email protected].

2567

on January 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

http://jb.asm.org/

MATERIALS AND METHODS

Bacterial strains, cloning vectors, and cultivation. S. natalensis ATCC 27448was routinely grown in YEME medium (25) without sucrose. Sporulation wasachieved in TBO medium (2). For pimaricin production, the strain was grown inYEME without sucrose. The same media were supplemented with thiostreptonwhen used for S. natalensis 6D82 growth and/or metabolite production. E. colistrain XL1-Blue MR (Stratagene) was used as a host for plasmid subcloning inplasmids pBluescript (Stratagene), pUC18, and pUC19. Candida utilis (synonym,Pichia jadinii) CECT 1061 was used for bioassay experiments. Phage KC515 (c�

attP::tsr::vph), a �C31-derived phage (38), was used for gene disruption experi-ments. Streptomyces lividans JII 1326 (17) served as a host for phage propagationand transfection. Infection with �6D8 (the KC515 recombinant derivative usedfor gene replacement) was carried out on R5 medium (25). Standard conditionsfor culture of Streptomyces species and isolation of phages were as described byKieser et al. (25).

Genetic procedures. Standard genetic techniques with E. coli and in vitro DNAmanipulations were as described by Sambrook and Russell (39). RecombinantDNA techniques in Streptomyces species and isolation of Streptomyces total andphage DNA were performed as previously described (25). Southern hybridiza-tion was carried out with probes labeled with digoxigenin by using the digoxige-nin DNA labeling kit (Roche Biochemicals).

DNA sequencing and analysis. Sequencing templates were obtained by ran-dom subcloning of fragments generated by controlled partial HaeIII digestions.DNA sequencing was accomplished by the dideoxynucleotide chain terminationmethod (40) with the Perkin Elmer Amplitaq dye-terminator sequencing systemon double-stranded DNA templates with an Applied Biosystems model 310sequencer (Foster City, Calif.). Each nucleotide was sequenced a minimum ofthree times on both strands. Alignment of sequence contigs was performed bythe DNAStar program Seqman (Madison, Wis.). DNA and protein sequenceswere analyzed with the EBI worldwide web BLAST and InterProScan servers aswell as with the CNRS Prodom Protein Domain Database server.

Construction of a pimR mutant. A 10,636-bp NotI fragment encompassing theentire pimR gene, pimK, and part of the pimS4 gene (Fig. 1) was cloned into aNotI-cut pBluescript vector to yield pNAF0. This plasmid was then digested withSalI and religated to yield pNAF1, which was used as a source of DNA for bothsequencing of pimR and for obtaining the DNA fragment used for gene replace-ment.

The pimR gene was disrupted by KC515 phage-mediated gene replacement asfollows. Plasmid pNAF1 was digested with NruI and religated to yield pNAF2.This treatment eliminates a 455-bp NruI fragment that codes for an internalpiece of PimR that contains an ATP/GTP-binding site (42) (see below) andresults in a mutant pimR gene with a frameshift beyond the new NruI site. A 2-kb

SacI-BglII fragment internal to the mutant pimR sequence was cloned into theSacI-BamHI sites of KC515 (38). Transfection of S. lividans protoplasts (25)resulted in a number of phage plaques that were screened by Southern hybrid-ization for the presence of pimR-derived sequences. One of the recombinants,�6D8, was selected and used to infect S. natalensis, thus allowing the selectionfor lysogen formation. Lysogens were selected by thiostrepton resistance on R5medium. Gene replacement was sought by repeated rounds of nonselectivegrowth in liquid YEME medium without sucrose, and the loss of the phage wasconfirmed by genomic Southern hybridization.

Isolation of total RNA. S. natalensis ATCC 27448 and S. natalensis �pimR weregrown for 48 h in YEME medium without sucrose (stationary phase of growth),the cultures were then mixed with 1 volume of 40% glycerol, and mycelia wereharvested by centrifugation and immediately frozen by immersion in liquid ni-trogen. Frozen mycelium was then broken by shearing in a mortar, and thefrozen lysate was added to buffer RLT (Qiagen) in the presence of 1.5% �-mer-captoethanol. RNeasy mini spin columns were used for RNA isolation accordingto the manufacturer’s instructions. RNA preparations were treated with DNaseI (Promega) to eliminate possible chromosomal DNA contamination.

Gene expression analysis by RT-PCR. Transcription was studied by using theSuperScript one-step reverse transcriptase PCR (RT-PCR) system with PlatinumTaq DNA polymerase (Invitrogen), with 5 ng of total RNA as the template.Conditions were as follows: first strand cDNA synthesis, 45°C for 40 min fol-lowed by 94°C for 2 min; amplification, 28 or 33 cycles of 98°C for 15 s, 60 to 70°C(depending of the set of primers used) for 30 s, and 72°C for 1 min. Primers (18-to 24-mers) (Table 1) were designed to generate PCR products of approximately500 bp. Negative controls were carried out with each set of primers and PlatinumTaq DNA polymerase to confirm the absence of contaminating DNA in the RNApreparations. The identity of each amplified product was corroborated by directsequencing with one of the primers.

Assay of pimaricin production. To assay pimaricin in culture broths, 0.5 ml ofculture was extracted with 4 ml of butanol, and the organic phase was diluted inwater-saturated butanol to bring the absorbance at 319 nm into the range of 0.1to 0.4 U. Solutions of pure pimaricin (DSM, Delft, The Netherlands) were usedas controls. To confirm the identity of pimaricin, a UV-visible absorption spec-trum (absorption peaks at 319, 304, 291, and 281 nm) was routinely determinedon a Hitachi U-2001 spectrophotometer. The fungicidal activity of pimaricin wastested by bioassay with C. utilis CECT 1061 as the test organism. Quantitativedetermination of pimaricin was performed with a Waters 600 high-performanceliquid chromatographer (HPLC) with a diode array UV detector set at 304 nm,fitted with a �Bondapak RP-C18 column (10 �m; 3.9 by 300 mm). Elution waswith a gradient (1 ml/min) of 100% methanol (methanol concentration: 50%from 0 to 3 min, up to 90% from 3 to 12 min, 90% from 12 to 20 min, down to

FIG. 1. Pimaricin biosynthetic gene cluster. The left fringe of the gene cluster is indicated in more detail and includes pimR (in dark grey), theglycosyl transferase-encoding open reading frame (pimK) thought to be involved in the attachment of the mycosamine moiety during pimaricinbiosynthesis, and the last two polyketide synthase (PKS) genes (pimS3 and pimS4) involved in the construction of the pimaricinolide (3). Pointedboxes indicate directions of transcription. All NotI (N) sites are shown; the SalI (S) and NruI (Nr) sites involved in vector construction are alsoindicated.

2568 ANTON ET AL. J. BACTERIOL.


.org/D

ownloaded from

http://jb.asm.org/

50% from 20 to 25 min, 50% from 25 to 30 min). The retention time forpimaricin was 14.5 min.

Cholesterol oxidase assay. Cholesterol oxidase activity was assessed by a cou-pled system with catalase (Roche Biochemicals) in which the cholesterol oxidasecatalyzes the oxidation of cholesterol to 4-cholesten-3-one with the reduction ofoxygen to hydrogen peroxide, which in the presence of catalase oxidizes meth-anol to formaldehyde. The latter, in the presence of NH4

�, reacts with acetylacetone to yield a yellow lutidine dye that is measured at 405 nm. The reactionmixture, containing 50 mM ammonium phosphate, 1.4 M methanol, 1.1 U ofcatalase/�l, 16 mM acetyl acetone, 0.1 �g of cholesterol/�l, and 1.4 U of com-mercial cholesterol oxidase/�l, was incubated at 37°C for 60 min. To assaycholesterol oxidase extracellular activity, culture broths of S. natalensis wild-ypeor �pimR strains were substituted for the commercial cholesterol oxidase in theabove reaction mixture.

Nucleotide sequence accession number. The sequence reported here has beendeposited in the GenBank database under the accession number AJ585085.

RESULTS

Cloning of pimR. pimR was identified by genomic walking byusing an S. natalensis ATCC 27448 cosmid library (2) andDNA segments from a previously unidentified open readingframe (orfX) upstream from pimK (which encodes a putativemycosamine transferase gene) on the left end of the pimaricingene cluster (3). The gene was sequenced from plasmidpNAF1 (see Materials and Methods) and turned out to be

separated by only 95 bp from the 5� end of pimK, with noobvious terminators between them (Fig. 1). The initiating ATGcodon of pimR is preceded by the sequence GAGAGG, whichcould potentially act as a ribosomal binding site. pimR is 3,597bp long, with an overall codon usage pattern in good agree-ment with that of typical Streptomyces genes; however, it con-tains a number of codons that are rare in such a G�C-richorganism. The presence of two TTA codons could be of par-ticular interest, since their involvement in the regulation ofdifferentiation and secondary metabolism in Streptomyces hasbeen proposed (26).

The pimR gene product is a multidomain protein that par-tially resembles positive regulators of antibiotic biosyntheticpathways. Computer-assisted analysis of the pimR gene prod-uct (1,198 amino acids with an estimated Mr of 130,184)showed a very high sequence identity (86%) with the whole ofprotein PTER of Streptomyces avermitilis, a putative regulatoryprotein of 1,096 amino acid residues whose encoding gene wasfound within the pte gene cluster, which is involved in thebiosynthesis of the pentane filipin (23, 34). PimR is some 100amino acid residues larger than PTER, thus resembling aPTER-like protein with an extra trans-Reg-C domain (Inter-Pro no. IPR001867) in its N terminus (Fig. 2A). This domain

TABLE 1. Primers for RT-PCR

Primer Sequence (5�–3�) Description

PIMR1S CCCGTCACGCCGTCGCCACAC Forward primer for pimR (before deletion)PIMR1AS CGCGTCCGCTCGTACACCTCAA Reverse primer for pimR (before deletion)PIMR2S CGTCTCCTGCCGCTCCTAC Forward primer for pimR (after deletion)PIMR2AS GCCTCCTCGATCTGCCCGTTGTGA Reverse primer for pimR (after deletion)PIMKS TGCTCAATCCGCTGCTCGTGCT Forward primer for pimKPIMKAS TCATCCGCCGCGACAGACC Reverse primer for pimKPIMS4S CCGCCAAGCAGGAACCGATAG Forward primer for pimS4PIMS4AS AGTGAGACCAGGGAGGAGGAGCAT Reverse primer for pimS4PIMS3S ACATCCAGTGAAACCGTCGTC Forward primer for pimS3PIMS3AS CCCTCCAGGCCCAGCGTGTA Reverse primer for pimS3PIMS2S CAGAAGCTCCGCGACTACCTAAAG Forward primer for pimS2PIMS2AS CCCACGGCGCCGAAGAAGTA Reverse primer for pimS2PIMIS GCGGTTCGGCCTCTTTCTTC Forward primer for pimIPIMIAS GCGCAGCAGCTCCTCGTC Reverse primer for pimIPIMJS CTGGTGTCCGAACTGTCCTT Forward primer for pimJPIMJAS CGCCGGTCCCGATCACATAGT Reverse primer for pimJPIMAS CCGCCGGGTGCTGGACTTCTC Forward primer for pimAPIMAAS CGCCGGTCCCGATCACATAGT Reverse primer for pimAPIMBS CTTCGCCGCCGCCTCCCTCCTCAC Forward primer for pimBPIMBAS CGCCGACGACCGCCACGACGACAT Reverse primer for pimBPIMES ACCACGATCACCTCCGCCCCTCAT Forward primer for pimEPIMEAS CCCGCCTTGCCCGCCTGCTC Reverse primer for pimEPIMCS GACGGCCGGGAGCTGGAGTATGTG Forward primer for pimCPIMCAS CTCGCCCGCCGTGATGATTTTGTT Reverse primer for pimCPIMGS CGACGGCAAACGGGTATGG Forward primer for pimGPIMGAS TCGGTGGAGGAGCGGAGGGAGAC Reverse primer for pimGPIMFS CGAGCCCGGCGCCGACCAGTA Forward primer for pimFPIMFAS GGCGGCCGACAGGACGAGAAG Reverse primer for pimFPIMS0S GGAATCGCCCCCGACGCAGTGG Forward primer for pimS0PIMS0AS CCGGACGAGGGGACGAGAACGA Reverse primer for pimS0PIMS1S ATGTCGAACGAGGAGAAGCTGC Forward primer for pimS1PIMS1AS CCGGAGGCGATGCTGATGTGATTG Reverse primer for pimS1PIMDS AACCGCCCAAAATGCTGAAACTGA Forward primer for pimDPIMDAS GCTCGGCCCGCTTGTGCTC Reverse primer for pimDPIMHS GTCCTCCACCTCACCGGCAACAG Forward primer for pimHPIMHAS CGATGAGCGTGATGGGCAGAAG Reverse primer for pimHLYSAS CGCCCGCCCACAGCAGGTCTTC Forward primer for lysALYSAAS TGGGGGTGCATGAGGAACTGAT Reverse primer for lysA

VOL. 186, 2004 REGULATION OF PIMARICIN BIOSYNTHESIS 2569


.org/D

ownloaded from

http://jb.asm.org/

(amino acids 34 to 98) resembles the DNA-binding domain atthe C terminus of the E. coli activator OmpR (29) and isusually found at the N termini of a family of antibiotic activa-tors known as SARPs (43).

No end-to-end counterparts were found in the protein da-tabases, however PimR showed significant sequence similaritywith several pathway specific regulators in its N-terminal 300amino acid residues (Fig. 2B). The highest scores were with theS. ambofaciens transcriptional activator SAR (40.3% identity)(18); AknO from the aclacinomycins pathway in S. galilaeus(40.2% identity) (37); Gra-orf9 from the granaticin pathway inS. violaceoruber (39.7% identity) (22); TylT and TylS from thetylosin pathway in S. fradiae (39% and 37.7% identity, respec-tively) (7), and ActII-orf4 from the actinorhodin pathway in S.coelicolor (36.5% identity) (21). All antibiotic activators de-scribed above belong to the family of Streptomyces antibioticregulatory proteins (SARPs) that appear to turn on the ex-pression of at least some of the genes of their respective clus-ters, thus controlling antibiotic production. The members ofthis family (typically between 277 and 666 amino acids) display

two conserved amino acid regions (Fig. 2B), one at the Nterminus with a proposed function in DNA binding (Trans-Reg-C domain; see above), and the second one at the C-terminal half of the protein. This later hypothetical domain isknown as bacterial transcriptional activator domain (BAD;ProDom PD005851), and it is also found in several transcrip-tional activators of Mycobacterium.

C-terminal of this region, PimR displays an AAA domain(InterPro no. IPR003593) that spans to complete the N-termi-nal half of the protein (amino acids 439 to 627) (Fig. 2A). Thisdomain is characteristic of a large family of ATPases associ-ated with diverse cellular activities, including cell cycle regu-lation, protein degradation, and protein transport, among oth-ers (35), and contains the ATP/GTP-binding or Walker Amotif A/G-X4-G-K-S/T (amino acids 447 to 454) (Fig. 2C)(42).

Comparison of database proteins with a stretch of 450 aminoacid residues around the AAA domain (amino acids 350 to800) yielded significant similarity with several proteins belong-ing to the LAL family of regulators (20), including various

FIG. 2. Domain structure and amino acid alignments of parts of the PimR protein. (A) Predicted domain structure of PimR. TRC, trans-reg-Cdomain (InterPro no. IPR001867) that resembles the DNA-binding domain of OmpR (see text); BAD, bacterial transcriptional activator domain(ProDom no. PD005851); AAA, domain characteristic of ATPases associated with diverse cellular activities (InterPro no. IPR003593). The asteriskindicates the location of the ATP/GTP-binding motif. Arrows indicate the regions of amino acid identity with proteins of the SARP, LAL, or CYCfamilies (see text). CYC, guanylate cyclase. (B) Alignment of the N-terminal portion of PimR with different proteins of the SARP family. Numbersindicate amino acid residues from the N terminus of the protein. Conserved amino acids are shown highlighted. SAR, Streptomyces ambofacienstranscriptional activator (SWALL Q9KWX4); AknO, SARP of the aclacinomycins pathway in Streptomyces galilaeus (SWALL Q9L4V0); Gra-Orf9, activator from the granaticin pathway in Streptomyces violaceoruber (SWALL Q9ZA48); TylS, positive regulator of tylosin biosynthesis inStreptomyces fradiae (SWALL Q9XCC4); ActII-Orf4, actinorhodin activator in S. coelicolor A3(2) (SWALL P46106). (C) Comparison of theWalker A (left) and B (right) motifs of PimR with those of other proteins. Numbers indicate amino acid residues from the N terminus of theprotein. Highly conserved amino acids are shown highlighted. Sc8d9.18, putative LAL regulatory protein from S. coelicolor A3(2) (SWALLQ9Z573); NysRI and NysRIII, LAL regulators of the nystatin pathway in S. noursei (SWALL Q9L4W1 and Q9L4V9); Sma0464 and Sma1591,putative guanylate cyclases encoded by plasmid pSymA of S. meliloti (SWALL Q930F6 and Q92YL0).



.org/D

ownloaded from

http://jb.asm.org/

putative transcriptional activators from Streptomyces coelicolor(products of the genes sc5f8.05c, sc8d9.18, and sc2g5.14c) (9)or the NysRIII and NysRI regulators of nystatin biosynthesis inS. noursei (12). The identity is restricted to the N-terminal 400amino acids of each of these proteins and therefore does notcover the characteristic C-terminal helix-turn-helix motif forDNA binding. Along with LAL regulators, BlastP homologysearches identified several guanylate cyclases, includingSma1591, Sma0464, and Sma1789 encoded by the megaplas-mid pSymA from Sinorhizobium meliloti (6) or Mll0576 fromMesorhizobium loti (24), although lacking the signature se-quence for these proteins (Prosite no. PDOC00425), which isusually located at the N terminus. The identity in this caseextends up to the C terminus of PimR (around 25% identitywith the whole of the cyclase except for the N-terminal 200amino acid residues).

Gene replacement of pimR. Since S. natalensis ATCC 27448has so far proved absolutely resistant to transformation byconventional procedures, we took advantage of the ability ofphage KC515, an attP-defective �C31 derivative (38), to infectS. natalensis to introduce DNA into this strain. The recombi-nant phage used for pimR inactivation, �6D8, was constructedas described in Materials and Methods and used to infect S.natalensis to obtain lysogens. Because phage KC515 and itsderivative lack attP, they can only form lysogens by homolo-gous recombination into the chromosome (Fig. 3A).

Five lysogens of S. natalensis were obtained by selection forthiostrepton resistance and tested for the lack of pimaricinproduction, since phage integration would disrupt the pimRtranscript. One of these mutants was randomly selected andnamed S. natalensis 6D82. The identity of the mutant wasconfirmed by Southern hybridization (Fig. 3B). This mutantwas then used to isolate thiostrepton-sensitive derivatives thathad undergone a second recombination event deleting theintegrated phage. These thiostrepton-sensitive isolates wereobtained after 6 rounds of nonselective growth in YEME me-dium. Of the nine colonies isolated, six were found to havereverted to the wild type while the other three harbored thedesired change.

One of these mutants, where pimR had been replaced by amutated version of it lacking the ATP-binding site, was ran-domly selected and named S. natalensis �pimR. ChromosomalDNAs isolated from S. natalensis ATCC 27448 and mutant�pimR and digested with both KpnI and PvuII were probedwith the 980-bp SmaI fragment used to construct the KC515derivative utilized for gene replacement (see Materials andMethods). Hybridizing bands of 2.4 and 0.8 kb were found forthe wild type as expected (Fig. 3B). However, in the disruptedmutant, a single 2.7-kb band was detected (Fig. 3B), indicatingthat a double crossover event had occurred. The observedhybridizing bands corresponded exactly to those expected ac-cording to the integration process depicted in Fig. 3. Figure 3also shows the hybridizing bands found for S. natalensis 6D82.

The new strain S. natalensis �pimR had growth and morpho-logical characteristics identical to those of the S. natalensis wildtype when grown on solid or liquid media, indicating that PimRhas no role in bacterial growth or differentiation.

Inactivation of pimR blocked pimaricin biosynthesis. Thefermentation broth produced by the mutant strain generatedby phage-mediated gene replacement, S. natalensis �pimR, was

extracted with butanol and analyzed for the presence of pima-ricin. Both the microbiological bioassay against C. utilis (datanot shown) and the HPLC assays indicated that no pimaricinwas being produced by the mutant strain �pimR (Fig. 4). Thisresult, together with the significant similarity of PimR in itsN-terminal section with well-known transcriptional activators,raised the question of which gene or genes were the potentialtarget of PimR activity.

Transcriptional control of pimaricin production. TotalRNA was prepared from the S. natalensis wild type and mutant�pimR after growth for 48 h (when pimaricin is actively pro-duced) (30) and used as a template for gene expression anal-

FIG. 3. Gene replacement of pimR. (A) Predicted restriction en-zyme polymorphism caused by gene replacement. The first crossoverevent is also indicated. The KpnI-PvuII restriction pattern before andafter replacement is shown. The probe is indicated by thick lines. Thefragment used for gene disruption derives from pNAF2 (see text fordetails). B, BamHI; K, KpnI; N, NruI; P, PvuII; S, SacI; Sm, SmaI; Z,BglII; wt, wild type. (B) Southern hybridization of the KpnI-PvuII-digested chromosomal DNA of the wild-type (lane 1), 6D82 (lane 2),and �pimR (lane 3) strains.



.org/D

ownloaded from

http://jb.asm.org/

ysis by RT-PCR. Primers for RT-PCR were specific to se-quences within pim genes (Table 1) and were designed toproduce cDNAs of approximately 500 bp. A primer pair de-signed to amplify a cDNA of the lysA gene (encoding diamin-opimelate decarboxylase) (Table 1) was used as an internalcontrol. Transcripts were analyzed from the 17 genes of thepim cluster, including pimR, after 28 PCR cycles. In the case ofpimR, transcripts were analyzed with pairs of primers locatedbefore and after the deletion. Whenever 28 cycles did not yielda product, analysis was repeated at 33 cycles. This analysis wascarried out at least three times for each primer pair.

All 17 genes were transcribed at 48 h in the S. natalensis wildtype; however, when we analyzed the transcription pattern in S.natalensis �pimR, we found virtually no transcripts for any ofthe genes of the cluster except for pimR (Fig. 5). No differencewas observed between the use of primer pairs located before orafter the deletion (Fig. 5A and B, respectively), thus indicatingthat transcription proceeded unabated across the site of theframeshift deletion in pimR. These results suggest that thegene replacement would have no polar effect on the transcrip-tion of pimK, which is located downstream from pimR. Thelack of transcripts of pimK in S. natalensis �pimR must there-fore be attributed to the absence of a functional PimR proteinin this strain and implies that pimK is transcribed from a

monocistronic transcript. Interestingly, the transcription pat-tern of the lysA gene was comparable in S. natalensis �pimRand in the parental strain (Fig. 5), thus validating the resultsdescribed above.

Increasing the number of PCR cycles from 28 to 33 allowedthe identification of transcripts for the remaining pim genesexcept for pimE (Fig. 5), thus suggesting that the mutant re-tains some transcription of the pim genes, albeit at very lowlevels. In any case, such low-level transcription was not suffi-cient for pimaricin production given that no pimaricin could bedetected in the culture broth of S. natalensis �pimR.

The strict control of PimR on the transcription of pimE wastotally unexpected for several reasons. First, due to the ab-sence of apparent transcriptional terminators in the short in-tergenic regions between pimA, pimB, pimE, pimC, pimG,pimF, and pimS0, these genes were thought to form an operonresulting in a transcript of more than 13,600 bp (3) which couldbe controlled coordinately. Second, no obvious function couldbe predicted for PimE in pimaricin biosynthesis, given thatPimE is a functional cholesterol oxidase (M. V. Mendes, N.Anton, J. F. Martín, and J. F. Aparicio, unpublished data).These results now could indicate that the above-indicatedgenes are actually transcribed from at least three differenttranscriptional units, namely pimAB, pimE, and pimCGFS0.

FIG. 4. (A) Comparison of HPLC analyses of butanol-extracted broths from S. natalensis wild-type (top) and �pimR (bottom) strains.Detection was carried out at A304. (B) Typical chromophore UV-visible absorption spectrum of pimaricin and its structure.



.org/D

ownloaded from

http://jb.asm.org/

However, in the absence of evidence indicating that all thetranscripts are equally stable, it is also possible that the mul-ticistronic transcript could be processed and subject to differ-ent rates of RNA degradation. Opposite to S. natalensis wild-type culture broths, no cholesterol oxidase activity could bedetected in S. natalensis �pimR cultures (data not shown), thusconfirming the lack of pimE transcription upon disruption ofpimR. The precise role of PimE on pimaricin biosynthesis isintriguing and requires further investigation.

DISCUSSION

Sequencing of the left-hand side of the pimaricin gene clus-ter revealed the presence of a gene, pimR, which could play arole as an activator for pimaricin biosynthesis in S. natalensis.Computer-assisted analysis of PimR revealed that it is a newmultidomain protein, with an N-terminal region strikingly sim-ilar to transcriptional activators of the SARP family (5, 43) anda C-terminal region with significant similarity with guanylatecyclases and transcriptional regulators of the LAL family(large ATP-binding regulators of the LuxR family) (20, 44),including the ATP/GTP-binding AAA domain present in theseprotein families but lacking the characteristic signature se-quence at the N terminus of guanylate cyclases or the LuxR-type helix-turn-helix motif for DNA binding present at the Cterminus of LAL regulators.

The significant similarity of the N-terminal section of PimR

with proteins of the SARP family raised two questions: first,whether PimR was actually a transcriptional activator, andsecond, whether, as an activator, it followed a regulatory be-havior comparable to that of SARPs. The absence of pimaricinproduction upon disruption of the gene provided the first in-dication of the functional role of PimR as an activator ofpimaricin biosynthesis, and the RT-PCR analysis revealed thatit controls the expression of all of the pimaricin genes except itsown transcription. SARPs activate transcription of their targetgenes by binding to heptameric repeats found around the �35regions of cognate promoters. These sequences, correspondingto the consensus 5�-TCGAGCG-3�, have been identified in thepromoter regions of actinorhodin (43), daunorubicin (41), andmithramycin (27) biosynthetic gene clusters; however, when weexamined the DNA sequence of the putative promoter regionsof the pimaricin genes, no such sequences were identified, thussuggesting that PimR probably follows a regulatory patterndifferent from those of SARPs. In fact, the presence of distinctWalker A and B motifs implicated in ATP/GTP binding inPimR suggests that its transcriptional activating activity is de-pendent on ATP/GTP hydrolysis.

The domain composition of PimR suggests that pimR couldhave been originated by merging a SARP-like gene with aguanylate cyclase gene that has lost its 5�-end 600 nucleotides.Why would S. natalensis develop such a chimera for the regu-lation of pimaricin production is unclear, but it is tempting tospeculate that the C-terminal region of PimR may act as a

FIG. 5. Gene expression analysis of the pimaricin gene cluster by RT-PCR. Analysis was carried out on S. natalensis wild-type (�) and �pimR(�) strains as indicated in Materials and Methods. The identity of each amplified product was corroborated by direct sequencing. The absence ofcontaminating DNA in the RNA samples was assessed by PCR. Due to its small size, transcription of pimF was assessed by using a reverse primer(Table 1) located within the coding sequence of the gene located downstream from it (i.e., pimS0). Transcription of pimR or its mutated versionwas assessed with two pairs of primers (Table 1), one pair located before the deletion (A) and the second pair located after the deletion (B).Transcription of the lysA gene (encoding diaminopimelate decarboxylase) was also assessed as an internal control. A diagram with the organizationof the genes within the pimaricin cluster and their putative transcripts is also included. The top panel shows the amplified products after 28 cyclesof PCR, and the bottom panel shows the same analysis after 33 cycles to detect low-level transcripts.



.org/D

ownloaded from

http://jb.asm.org/

sensory transducer for the SARP-like region to bind DNA andactivate transcription in an ATP-dependent manner. Bothguanylate cyclases and the prototype of LAL regulatory pro-teins, the regulator of the maltose regulon in E. coli, MalT(11), interact with effectors to modulate their activities (19),and that could be the case for PimR.

The control of pimaricin biosynthesis exerted by PimR takesplace through the specific transcriptional activation of all of thekey enzyme-encoding genes for pimaricinolide construction.PimR either controls multiple pimaricin biosynthetic promot-ers or activates another hierarchical regulator(s). In any case,this supposes an important energy savings given that transcriptsynthesis is energy dependent, a fact that acquires special rel-evance for the giant mRNAs that are supposed to direct thesynthesis of pimaricin (3).

The absence of pimE transcripts on S. natalensis �pimRcompared to those of its surrounding genes was unexpected.The only indication for a functional role of pimE in pimaricinbiosynthesis was its chromosomal location just in the middle ofthe pimaricin gene cluster (2, 3); however, the activity of itsgene product as a functional cholesterol oxidase (Mendes etal., unpublished) suggested that such a location could be amatter of chance. Now, these results strongly suggest thatPimE is actually involved in pimaricin biosynthesis and raisethe question of what is the precise mechanism that drives thecellular role of this enzyme. Interestingly, a cholesterol oxi-dase-encoding gene, pteG, has been also identified within thegene cluster of the 26-membered polyene filipin produced by S.avermitilis (23, 34). Further experimental analyses (now inprogress) will hopefully provide the answer to this questionand to the generality of this phenomenon.

Besides that of pimaricin (2, 3), no other biosynthetic genecluster for a small polyene has been characterized to date. Thefact that, as yet, the only orthologue of PimR is PTER, aputative regulator of the pentane filipin produced by S. aver-mitilis (23, 34), suggests that common regulatory circuits couldoperate for the expression of these kind of molecules and thatthey are not shared by other polyenes. The absence of similarregulatory proteins in the recently characterized nystatin (12),amphotericin (13), or candicidin (14) gene clusters supportssuch a hypothesis. Further characterization of the PimR pro-tein might prove useful for future applications to control andimprove the expression of genes involved in the biosynthesis ofsmall polyenes and thus increase their production.

ACKNOWLEDGMENTS

This work was supported by a grant from the Comision Interminis-terial de Ciencia y Tecnología (CICYT) to J.F.A. (BIO2001--0040).M.V.M. received a fellowship of the Fundacao para a Ciencia e aTecnologia (PRAXIS XXI/BD/15850/98). N.A. was the recipient of anF.P.U. fellowship from Ministerio de Educacion, Cultura, y Deporte(AP2002-1446).

We thank M. Driessen (DSM) and E. Recio for helpful discussionsand J. Merino and B. Martín for excellent technical assistance.

REFERENCES

1. Aparicio, J. F., I. Molnar, T. Schwecke, A. Konig, S. F. Haydock, L. E. Khaw,J. Staunton, and P. F. Leadlay. 1996. Organisation of the biosynthetic genecluster for rapamycin in Streptomyces hygroscopicus: analysis of the enzymaticdomains in the modular polyketide synthase. Gene 169:9–16.

2. Aparicio, J. F., A. J. Colina, E. Ceballos, and J. F. Martín. 1999. Thebiosynthetic gene cluster for the 26-membered ring polyene macrolide pima-ricin: a new polyketide synthase organization encoded by two subclustersseparated by functionalization genes. J. Biol. Chem. 274:10133–10139.

3. Aparicio, J. F., R. Fouces, M. V. Mendes, N. Olivera, and J. F. Martín. 2000.A complex multienzyme system encoded by five polyketide synthase genes isinvolved in the biosynthesis of the 26-membered polyene macrolide pimari-cin in Streptomyces natalensis. Chem. Biol. 7:895–905.

4. Aparicio, J. F., P. Caffrey, J. A. Gil, and S. B. Zotchev. 2003. Polyeneantibiotic biosynthesis gene clusters. Appl. Microbiol. Biotechnol. 61:179–188.

5. Arias, P., M. A. Fernandez-Moreno, and F. Malpartida. 1999. Characteriza-tion of the pathway-specific positive transcriptional regulator for actinor-hodin biosynthesis in Streptomyces coelicolor A3(2) as a DNA-binding pro-tein. J. Bacteriol. 181:6958–6968.

6. Barnett, M. J., R. F. Fisher, T. Jones, C. Komp, A. P. Abola, F. Barloy-Hubler, L. Bowser, D. Capela, F. Galibert, J. Gouzy, M. Gurjal, A. Hong, L.Huizar, R. W. Hyman, D. Kahn, M. L. Kahn, S. Kalman, D. H. Keating, C.Palm, M. C. Peck, R. Surzycki, D. H. Wells, K. C. Yeh, R. W. Davis, N. A.Federspiel, and S. R. Long. 2001. Nucleotide sequence and predicted func-tions of the entire Sinorhizobium meliloti pSymA megaplasmid. Proc. Natl.Acad. Sci. USA 98:9883–9888.

7. Bate, N., A. R. Butler, A. R. Gandecha, and E. Cundliffe. 1999. Multipleregulatory genes in the tylosin biosynthetic cluster of Streptomyces fradiae.Chem. Biol. 6:617–624.

8. Bate, N., G. Stratigopoulos, and E. Cundliffe. 2002. Differential roles of twoSARP-encoding regulatory genes during tylosin biosynthesis. Mol. Micro-biol. 43:449–458.

9. Bentley, S. D., K. F. Chater, A. M. Cerdeno-Tarraga, G. L. Challis, N. R.Thomson, K. D. James, D. E. Harris, M. A. Quail, H. Kieser, D. Harper, A.Bateman, S. Brown, G. Chandra, C. W. Chen, M. Collins, A. Cronin, A.Fraser, A. Goble, J. Hidalgo, T. Hornsby, S. Howarth, C. H. Huang, T.Kieser, L. Larke, L. Murphy, K. Oliver, S. O’Neil, E. Rabbinowitsch, M. A.Rajandream, K. Rutherford, S. Rutter, K. Seeger, D. Saunders, S. Sharp, R.Squares, S. Squares, K. Taylor, T. Warren, A. Wietzorrek, J. Woodward,B. G. Barrell, J. Parkhill, and D. A. Hopwood. 2002. Complete genomesequence of the model actinomycete Streptomyces coelicolor A3(2). Nature417:141–147.

10. Bibb, M. 1996. The regulation of antibiotic production in Streptomyces coeli-color A3(2). Microbiology 142:1335–1344.

11. Boos, W., and H. Shuman. 1998. Maltose/maltodextrin system of Escherichiacoli: transport, metabolism, and regulation. Microbiol. Mol. Biol. Rev. 62:204–229.

12. Brautaset, T., O. N. Sekurova, H. Sletta, T. E. Ellingsen, A. R. Strom, S.Valla, and S. B. Zotchev. 2000. Biosynthesis of the polyene antifungal anti-biotic nystatin in Streptomyces noursei ATCC 11455: analysis of the genecluster and deduction of the biosynthetic pathway. Chem. Biol. 7:395–403.

13. Caffrey, P., S. Lynch, E. Flood, S. Finnan, and M. Oliynyk. 2001. Ampho-tericin biosynthesis in Streptomyces nodosus: deductions from analysis ofpolyketide synthase and late genes. Chem. Biol. 8:713–723.

14. Campelo, A. B., and J. A. Gil. 2002. The candicidin gene cluster fromStreptomyces griseus IMRU 3570. Microbiology 148:51–59.

15. Champness, W. C., and K. F. Chater. 1994. Regulation and integration ofantibiotic production and morphological differentiation in Streptomyces spp.,p. 61–93. In P. Piggot, C. P. Moran, and P. Youngman (ed.), Regulation ofbacterial differentiation. American Society for Microbiology, Washington,D.C.

16. Chater, K. F., and M. Bibb. 1997. Regulation of bacterial antibiotic produc-tion, p. 59–105. In H. Kleinkauf and H. von Doren (ed.), Biotechnology, vol.7. Products of secondary metabolism. VCH, Weinheim, Germany.

17. Chater, K. F., C. J. Bruton, and J. E. Suarez. 1981. Restriction mapping ofthe DNA of the Streptomyces temperate phage �C31 and its derivatives.Gene 14:183–194.

18. Culebras, E., E. Martínez, A. Carnero, and F. Malpartida. 1999. Cloning andcharacterization of a regulatory gene of the SARP family and its flankingregion from Streptomyces ambofaciens. Mol. Gen. Genet. 262:730–737.

19. Danot, O. 2001. A complex signaling module governs the activity of MalT,the prototype of an emerging transactivator family. Proc. Natl. Acad. Sci.USA 98:435–440.

20. De Schrijver, A., and R. De Mot. 1999. A subfamily of MalT-related ATP-dependent regulators in the LuxR family. Microbiology 145:1287–1288.

21. Fernandez-Moreno, M. A., J. L. Caballero, D. A. Hopwood, and F. Malpar-tida. 1991. The act cluster contains regulatory and antibiotic export genes,direct targets for translational control by the bldA transfer RNA gene ofStreptomyces. Cell 66:769–780.

22. Ichinose, K., D. J. Bedford, D. Tornus, A. Bechthold, M. J. Bibb, W. P. Revill,H. G. Floss, and D. A. Hopwood. 1998. The granaticin biosynthetic genecluster of Streptomyces violaceoruber Tu22: sequence analysis and expressionin a heterologous host. Chem. Biol. 5:647–659.

23. Ikeda, H., J. Ishikawa, A. Hanamoto, M. Shinose, H. Kikuchi, T. Shiba, Y.Sakaki, M. Hattori, and S. Omura. 2003. Complete genome sequence andcomparative analysis of the industrial microorganism Streptomyces avermiti-lis. Nat. Biotechnol. 21:526–531.

24. Kaneko, T., Y. Nakamura, S. Sato, E. Asamizu, T. Kato, S. Sasamoto, A.Watanabe, K. Idesawa, A. Ishikawa, K. Kawashima, T. Kimura, Y. Kishida,C. Kiyokawa, M. Kohara, M. Matsumoto, A. Matsuno, Y. Mochizuki, S.



.org/D

ownloaded from

http://jb.asm.org/

Nakayama, N. Nakazaki, S. Shimpo, M. Sugimoto, C. Takeuchi, M. Yamada,and S. Tabata. 2000. Complete genome structure of the nitrogen-fixingsymbiotic bacterium Mesorhizobium loti. DNA Res. 7:331–338.

25. Kieser, T., M. J. Bibb, M. J. Buttner, K. F. Chater, and D. A. Hopwood. 2000.Practical Streptomyces genetics. John Innes Foundation, Norwich, UnitedKingdom.

26. Leskiw, B. K., E. J. Lawlor, J. M. Fernandez-Abalos, and K. F. Chater. 1991.TTA codons in some genes prevent their expression in a class of develop-mental, antibiotic-negative, Streptomyces mutants. Proc. Natl. Acad. Sci.USA 88:2461–2465.

27. Lombo, F., A. F. Brana, C. Mendez, and J. A. Salas. 1999. The mithramycingene cluster of Streptomyces argillaceus contains a positive regulatory geneand two repeated DNA sequences that are located at both ends of thecluster. J. Bacteriol. 181:642–647.

28. Madduri, K., and C. R. Hutchinson. 1995. Functional characterization andtranscriptional analysis of the dnrR1 locus, which controls daunorubicinbiosynthesis in Streptomyces peucetius. J. Bacteriol. 177:1208–1215.

29. Martinez-Hackert, E., and A. M. Stock. 1997. Structural relationships in theOmpR family of winged-helix transcription factors. J. Mol. Biol. 269:301–312.

30. Mendes, M. V., E. Recio, R. Fouces, R. Luiten, J. F. Martin, and J. F.Aparicio. 2001. Engineered biosynthesis of novel polyenes: a pimaricin de-rivative produced by targeted gene disruption in Streptomyces natalensis.Chem. Biol. 8:635–644.

31. Mizuno, T., and I. Tanaka. 1997. Structure of the DNA-binding domain ofthe OmpR family of response regulators. Mol. Microbiol. 24:665–667.

32. Molnar, I., J. F. Aparicio, S. F. Haydock, L. E. Khaw, T. Schwecke, A. Konig,J. Staunton, and P. F. Leadlay. 1996. Organisation of the biosynthetic genecluster for rapamycin in Streptomyces hygroscopicus: analysis of genes flank-ing the polyketide synthase. Gene 169:1–7.

33. Narva, K. E., and J. S. Feitelson. 1990. Nucelotide sequence and transcrip-tional analysis of the redD locus of Streptomyces coelicolor A3(2). J. Bacteriol.172:326–333.

34. Omura, S., H. Ikeda, J. Ishikawa, A. Hanamoto, C. Takahashi, M. Shinose,Y. Takahashi, H. Horikawa, H. Nakazawa, T. Osonoe, H. Kikuchi, T. Shiba,

Y. Sakaki, and M. Hattori. 2001. Genome sequence of an industrial micro-organism Streptomyces avermitilis: deducing the ability of producing second-ary metabolites. Proc. Natl. Acad. Sci. USA 98:12215–12220.

35. Patel, S., and M. Latterich. 1998. The AAA team: related ATPases withdiverse functions. Trends Cell Biol. 8:65–71.

36. Perez-Llarena, F. J., P. Liras, A. Rodríguez-García, and J. F. Martín. 1997.A regulatory gene (ccaR) required for cephamycin and clavulanic acid pro-duction in Streptomyces clavuligerus: amplification results in overproductionof both �-lactam compounds. J. Bacteriol. 179:2053–2059.

37. Raty, K., T. Kunnary, J. Hakala, P. Mantsala, and K. Ylihonko. 2000. A genecluster from Streptomyces galilaeus involved in glycosylation of aclarubicin.Mol. Gen. Genet. 264:164–172.

38. Rodicio, M. R., C. J. Bruton, and K. F. Chater. 1985. New derivatives of theStreptomyces temperate phage �C31 useful for the cloning and functionalanalysis of Streptomyces DNA. Gene 34:283–292.

39. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratorymanual, 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor,N.Y.

40. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chainterminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463–5467.

41. Tang, L., A. Grimm, Y. X. Zhang, and C. R. Hutchinson. 1996. Purificationand characterization of the DNA-binding protein DnrI, a transcriptionalfactor of daunorubicin biosynthesis in Streptomyces peucetius. Mol. Micro-biol. 22:801–813.

42. Walker, J. E., M. Saraste, M. J. Runswick, and N. J. Gay. 1982. Distantlyrelated sequences in the alpha- and beta-subunits of ATP synthase, myosin,kinases and other ATP-requiring enzymes and a common nucleotide bindingfold. EMBO J. 1:945–951.

43. Wietzorrek, A., and M. Bibb. 1997. A novel family of proteins that regulatesantibiotic production in streptomycetes appears to contain an OmpR-likeDNA-binding fold. Mol. Microbiol. 25:1177–1184.

44. Wilson, D. J., Y. Xue, K. E. Reynolds, and D. H. Sherman. 2001. Charac-terization and analysis of the PikD regulatory factor in the pikromycinbiosynthetic pathway of Streptomyces venezuelae. J. Bacteriol. 183:3468–3475.



.org/D

ownloaded from

http://jb.asm.org/

Identiﬁcation of PimR as a Positive Regulator of Pimaricin ... · rapamycin pathway in...

Documents

Transcript of Identiﬁcation of PimR as a Positive Regulator of Pimaricin ... · rapamycin pathway in...