Identification and Utility of FdmR1 as a Streptomyces Antibiotic ...

JOURNAL OF BACTERIOLOGY, Aug. 2008, p. 5587–5596 0021-9193/08/$08.00⫹0 doi:10.1128/JB.00592-08 Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Vol. 190, No. 16

Identification and Utility of FdmR1 as a Streptomyces Antibiotic Regulatory Protein Activator for Fredericamycin Production in Streptomyces griseus ATCC 49344 and Heterologous Hosts䌤† Yihua Chen,1 Evelyn Wendt-Pienkowski,1 and Ben Shen1,2,3* Division of Pharmaceutical Sciences,1 University of Wisconsin National Cooperative Drug Discovery Group,2 and Department of Chemistry,3 University of Wisconsin-Madison, Madison, Wisconsin 53705-2222 Received 29 April 2008/Accepted 6 June 2008

The fredericamycin (FDM) A biosynthetic gene cluster, cloned previously from Streptomyces griseus ATCC 49344, contains three putative regulatory genes, fdmR, fdmR1, and fdmR2. Their deduced gene products show high similarity to members of the Streptomyces antibiotic regulatory protein (SARP) family (FdmR1) or to MarR-like regulators (FdmR and FdmR2). Here we provide experimental data supporting FdmR1 as a SARP-type activator. Inactivation of fdmR1 abolished FDM biosynthesis, and FDM production could be restored to the fdmR1::aac(3)IV mutant by expressing fdmR1 in trans. Reverse transcription-PCR transcriptional analyses revealed that up to 26 of the 28 genes within the fdm gene cluster, with the exception of fdmR and fdmT2, were under the positive control of FdmR1, directly or indirectly. Overexpression of fdmR1 in S. griseus improved the FDM titer 5.6-fold (to about 1.36 g/liter) relative to that of wild-type S. griseus. Cloning of the complete fdm cluster into an integrative plasmid and subsequent expression in heterologous hosts revealed that considerable amounts of FDMs could be produced in Streptomyces albus but not in Streptomyces lividans. However, the S. lividans host could be engineered to produce FDMs via constitutive expression of fdmR1; FDM production in S. lividans could be enhanced further by overexpressing fdmC, encoding a putative ketoreductase, concomitantly with fdmR1. Taken together, these studies demonstrate the viability of engineering FDM biosynthesis and improving FDM titers in both the native producer S. griseus and heterologous hosts, such as S. albus and S. lividans. The approach taken capitalizes on FdmR1, a key activator of the FDM biosynthetic machinery. pathways calls for knowledge of large gene cluster transfer, special precursor complementation, posttranslational modifications of related proteins, and appropriate host and medium selection. Effective heterologous expression also calls for a rigorous understanding of the regulatory systems involved in secondary metabolism (15, 40, 45). Decades of persistent effort involving several Streptomyces model strains have unveiled the regulatory systems for some representative biosynthetic pathways, and a number of regulators responsible for initiating secondary metabolite production have been identified (6). Conclusively, the regulation of biosynthetic activation is regarded as taking place genetically at several levels. The first level operates via genes that dictate global regulators responsible for both morphological differentiation and secondary metabolite production. This level of regulation can be exemplified by bldA, which encodes a tRNA recognizing the rare leucine codon UUA. Inactivation of bldA leads to developmental defects in aerial hyphae and in spore formation and to abolishment of actinorhodin and undecylprodigiosin production in Streptomyces coelicolor A3(2) (8). The next level of activation involves pleiotropic regulators controlling biosynthetic pathways for more than one secondary metabolite. For instance, AfsR controls the production of actinorhodin, undecylprodigiosin, and calcium-dependent antibiotic in S. coelicolor A3(2) (16). The third level of regulation refers to pathway-specific regulators, affecting only one biosynthetic pathway. These regulators have been classified into different families, such as the LAL family (large ATP-binding regulators of the LuxR family) (43), the LysR-like regulators (10),

Members of the genus Streptomyces are gram-positive filamentous bacteria that continue to be a prolific source of bioactive secondary metabolites, including many clinically important antimicrobial and anticancer drugs as well as agents with agricultural or veterinary applications (7). These metabolites are predominantly products of complex biosynthetic pathways, typically activated in a growth-phase-dependent manner coinciding with the formation of aerial mycelia in solid media or confining to stationary phase in liquid cultures. The production of these natural products is often controlled by subtle and precise regulatory systems (6). A fundamental comprehension of these regulatory systems is undoubtedly helpful in understanding biosynthetic transformations, thereby enhancing opportunities for combinatorial biosynthesis to afford new compounds, and in optimizing metabolite titers in a strategic manner. Recently, heterologous expression has emerged as a powerful tool to investigate secondary metabolite biosynthetic pathways. This approach is especially advantageous for those clusters from organisms recalcitrant to genetic manipulations or resistant to effective culturing within a laboratory setting. The successful heterologous expression of defined biosynthetic * Corresponding author. Mailing address: Division of Pharmaceutical Sciences, School of Pharmacy, University of Wisconsin-Madison, 777 Highland Ave., Madison, WI 53705-2222. Phone: (608) 263-2673. Fax: (608) 262-5345. E-mail: [email protected]. † Supplemental material for this article may be found at http://jb .asm.org/. 䌤 Published ahead of print on 13 June 2008. 5587

5588

CHEN ET AL.

and the Streptomyces antibiotic regulatory protein (SARP) family (42). The SARP family not only contains pathway-specific regulators but is also composed of pleiotropic regulatory proteins, such as AfsR (34). Members of this family are all activators characterized by their N-terminal DNA-binding domain, resembling the DNA-binding domain of OmpR (26, 42), and by an accompanying bacterial transcriptional activation domain (44). The SARP regulators vary in length from fewer than 300 residues (e.g., DnrI has 272 amino acids, containing only the DNA-binding and transcriptional activation domains) (35) to about 1,000 residues. For instance, AfsR, which is composed of 993 amino acids, contains an ATPase domain and a tetratricopeptide repeat domain in addition to the DNA-binding and transcriptional activation domains (34). These regulators recognize direct repeats within the promoter regions of the genes being regulated, and at least some of them bind to heptameric direct repeats (TCGAGXX) spaced 4 nucleotides (nt) or 15 nt from each other and located stringently 8 nt upstream of the ⫺10 regions (2, 32, 42). After binding to the specific direct repeats, the SARP regulators are proposed to initiate transcription of controlled genes by recruitment of RNA polymerase to the appropriate sites (34). Fredericamycin (FDM) A is an aromatic pentadecaketide containing a unique asymmetric carbaspirocyclic skeleton, which is produced together with intermediates FDM C, B, and E by Streptomyces griseus ATCC 49344 (Fig. 1a) (9, 27, 33). Previous work has shown that FDM A possesses moderate antitumor bioactivity and displays cytotoxicity against several cell lines (9, 38). These activities might result from inhibition of topoisomerases I and II (19) or the peptidyl-prolyl cis-trans isomerase Pin1, a critical regulator of mitosis and oncogenesis (23; K. P. Lu and G. Fisher, January 2004, PCT international patent application WO2004002429). Given that Pin1 has become an important target of anticancer drug discovery and that FDM A distinguishes itself from all other studied Pin1 inhibitors by its unique structural skeleton, FDM A represents a new chemotype for anticancer drug leads (20). Therefore, a better understanding of the FDM biosynthetic regulatory system and increasing the FDM A titer by regulatory pathway manipulations should significantly hasten not only anticancer drug discovery efforts but also other basic research on this fascinating small-molecule natural product. Prior to this study, all that was known about the regulatory system controlling FDM production was that there were three postulated regulatory genes (fdmR, fdmR1, and fdmR2) located within the fdm cluster (Fig. 1b) (39). Herein, we demonstrate that one of the putative regulatory genes, fdmR1, encodes a SARP-type activator that controls, directly or indirectly, up to 26 of the 28 genes within the fdm cluster, with the exception of fdmR and fdmT2, and that the FDM titer can be improved by a factor of 5.6 (to about 1.36 g/liter) by overexpressing fdmR1 in the wild-type S. griseus strain. The fdm gene cluster was successfully expressed previously in a high-copy-number plasmid in Streptomyces albus J1074 (39). We now show that when the fdm cluster was transferred to an integrative plasmid, it could also be expressed efficiently in S. albus. In contrast, the fdm cluster was not expressed in Streptomyces lividans K4-114, and fdm expression in S. lividans was realized only upon constitutive coexpression

J. BACTERIOL.

of fdmR1. The FDM titer from the S. lividans recombinant strain was finally optimized by tandem expression of fdmR1 and fdmC, which encodes a putative ketoreductase. MATERIALS AND METHODS Bacterial strains, plasmids, and growth conditions. Escherichia coli DH5␣ (30) and ET12567/pUZ8002 (24) were grown in Luria-Bertani broth (30). S. griseus ATCC 49344 (9, 39), S. albus J1074 (39), and S. lividans K4-114 (46) were cultured in R2YE medium (17) for seed culture preparation. Antibiotic producing medium (APM) was used as the fermentation medium for FDM production (39). For protoplast preparation, the Streptomyces strains were grown in YEME medium (17). MYME (17) and APM (39) were used for genomic DNA and total RNA isolation, respectively. pGEM7zf and pSP72 were from Promega (Madison, WI), and pANT841 (11), pHJL401 (17), pSET151 (17), pSET152 (17), pWHM3 (17), and pWHM1250 (25) were described previously. The concentrations of antibiotics used were 100 ␮g/ml (ampicillin), 50 ␮g/ml (apramycin [Am]), and 10 ␮g/ml (thiostrepton) (17, 30). DNA manipulation and sequence analysis. General DNA manipulations were performed as described previously (30). PCRs were performed with a GoTaq DNA polymerase kit (Promega, Madison, WI). S. griseus genomic DNA was isolated by a standard procedure (17). Protoplast preparation and transformation were performed as described previously (17). Southern analysis was performed with a DIG High Prime labeling kit according to the manufacturer’s protocol (Roche Biochemicals, Nutley, NJ). Homologous sequence database searching and multiple alignments were executed with BLASTP and ClustalX, respectively. Protein secondary structure analysis was performed with PredictProtein (29). Overexpression of fdmR1 in the S. griseus wild-type strain. A double-stranded linker containing an upstream NsiI end, an XbaI site, the native DNA sequence of a portion of the cluster, starting just upstream of the probable ribosomal binding site and ending with the PstI site within fdmR1, and a SacI downstream end was created with two oligonucleotides (5⬘-TTCTAGATGCAAACCAACG GGAGGGCCAGTGCTGCAGGAGCT-3⬘ and 5⬘-CCTGCAGCACTGGCCCT CCCGTTGGTTTGCATCTAGAATGCA-3⬘). This linker was ligated into pGEM7zf at the NsiI/SacI sites. A 2.4-kb PstI/BglII fragment containing the remainder of the fdmR1 gene was then inserted into the PstI/BamHI sites to construct the complete fdmR1 gene. The recreated fdmR1 gene was removed as a 3.4-kb XbaI/HindIII fragment and inserted at the same sites in pWHM1250 behind the ermE* promoter to yield pBS4045. Introduction of pBS4045 into the S. griseus wild-type strain via protoplast transformation finally afforded the recombinant strain SB4012, in which the expression of fdmR1 is under the control of the constitutive ermE* promoter (17). Inactivation of fdmR1 by gene replacement and complementation of the S. griseus fdmR1::aac(3)IV mutant by fdmR1 expression in trans. A 3.3-kb BglII/StuI fragment containing fdmR1 was isolated and cloned into the BamHI/EcoRV sites of pSP72. The aac(3)IV Am resistance gene cassette (17) was then cloned as a ClaI/KpnI fragment into the same sites within fdmR1, replacing a 790-bp internal fragment of the gene. The mutated fdmR1 gene was then moved as a 3.8-kb BglII/SphI fragment into the same sites of pSET151 to generate the gene replacement construct pBS4046. After introduction of pBS4046 into the wild-type S. griseus strain via E. coli-S. griseus conjugation, Am-resistant transformants were screened for a thiostrepton-sensitive phenotype to identify double-crossover mutants. The genotype of the double-crossover mutant strain SB4013 was confirmed by Southern analysis, using the 0.48-kb PstI/ClaI fragment inside fdmR1 as the probe (see Fig. S1 in the supplemental material). Plasmid pBS4047 was constructed by removing the 2.7-kb EcoRI/HindIII fragment containing both the ermE* promoter and the fdmR1 gene from pBS4045 and inserting it into the same sites of the middle-copy-number plasmid pHJL401. Introduction of pBS4047 into SB4013 via protoplast transformation afforded the SB4014 strain, in which the fdmR::aac(3)IV mutation was complemented by ermE*-controlled fdmR1 expression in trans. Construction of fdm gene cluster expression construct and production of FDM in heterologous hosts. The 28.1-kb EcoRI/XbaI fragment containing the whole fdm cluster was excised from pBS4028 and inserted into the same sites of pSET152 to generate the integrative plasmid pBS4048. For FDM production in S. albus and S. lividans, the pBS4048 plasmid was introduced into S. albus J1074 and S. lividans K4-114 by E. coli-Streptomyces conjugation to afford the S. albus SB4015 and S. lividans SB4016 recombinant strains, respectively. To upregulate the expression of the fdm cluster in heterologous hosts, the 2.9-kb NcoI/BglII fragment containing fdmR1 with its native promoter was excised from pBS4028, cloned into the same sites of pANT841, and then moved as

VOL. 190, 2008

FdmR1 AS A SARP ACTIVATOR

5589

FIG. 1. (a) Proposed biosynthetic pathway for FDM. (b) Map of fdm biosynthetic gene cluster. Regulatory genes are filled with black. Other genes selected for RT-PCR analysis are marked with striped lines. Putative promoters containing SARP binding sites are marked with solidly lined arrows. Other postulated promoters are indicated with dashed line arrows. (c) Alignment of intergenic promoter regions containing putative SARP binding sites inside the fdm cluster. The heptameric repeat sequences are indicated by solid underlines, and the ⫺10 regions are marked with dashed underlines. The arrow indicates the direction of transcription.

a 3.0-kb EcoRI/HindIII fragment into the same sites of the high-copy-number plasmid pWHM3 to yield pBS4049. Introduction of pBS4049 into SB4016 by protoplast transformation afforded the recombinant strain SB4017, in which the copy number of fdmR1 was significantly increased. Similarly, SB4019 was obtained by introducing pBS4045 into SB4016, in which the expression of the fdm cluster in S. lividans is ensured by ermE*-controlled constitutive expression of fdmR1 in trans. To optimize FDM production in S. lividans, a 2.6-kb PstI/SphI fragment containing fdmC was first excised from pBS4028 and cloned into the same site of pWHM1250 to generate pBS4050, in which the expression of fdmC is under the control of ermE*. A 2.3-kb BamHI/XmnI fragment, containing the fdmR1 gene, was then cloned from pBS4045 and inserted into BamHI/SnaBI sites of pBS4050 to afford pBS4051, in which the expression of both fdmC and fdmR1 is under the

control of ermE*. Introduction of pBS4050 or pBS4051 into SB4016 by protoplast transformation yielded SB4019 or SB4020, respectively, in which the expression of the fdm cluster in S. lividans is augmented by the ermE*-controlled constitutive expression of fdmC alone (SB4019) or fdmC and fdmR1 concomitantly (SB4020). RT-PCR. The total RNAs of different Streptomyces strains were isolated from mycelia (100 mg for every strain) which had been growing for 3 days in APM. After motorized grinding with liquid nitrogen, powders of mycelia were subjected to RNA extraction using an RNeasy Mini kit (Qiagen, Valencia, CA). The isolated total RNAs were then treated with RNase-Free DNase (Qiagen, Valencia, CA) to remove possible contaminant DNA. Complete digestion of the contaminant DNA was confirmed by PCR with the primers for fdmR1 or fdmO, using GoTaq DNA polymerase (see Fig. S2 in the supplemental material). RNA

5590

CHEN ET AL.

J. BACTERIOL.

FIG. 2. (a) Multiple alignment of the FdmR1 N-terminal sequence with the N-terminal sequence of RubS and the well-characterized SARP activators ActII-ORF4 and DnrI. The predicted secondary structure of the FdmR1 N-terminal DNA-binding domain, including the HTH variant, is marked. The transcriptional activation domain is indicated with a gray bar. (b) Multiple alignment of the FdmR1 C-terminal sequence with homologous regions of AfsR, RubS, and SnorA. The ATP binding motifs are marked with black squares.

concentrations were determined by measuring the UV absorbance at 260 nm (30). The primers used for reverse transcription-PCR (RT-PCR) were as follows: for FdmC, 5⬘-GCTTCTGCTCGCGTTCCGCCG-3⬘ and 5⬘-CCGCCGTCGATC TGCACCACC-3⬘; for FdmF, 5⬘-CGGCCACACGGCGCATCACC-3⬘ and 5⬘-C CGGCACGTAGTCGAGGTCGC-3⬘; for FdmL, 5⬘-CGCCGTCACCTACCGC GTCG-3⬘ and 5⬘-CGCCGAGCACGGACATGTGGG-3⬘; for FdmO, 5⬘-CTGG TCACCGGCGGTACCCG-3⬘ and 5⬘-CCCGTCGACATGCAGGGTCTCG-3⬘; for FdmR, 5⬘-CCGGCAGTCCCGGCGCAG-3⬘ and 5⬘-GCCGTGGTCCGCCT CGTCC-3⬘; for FdmR1, 5⬘-CTCGCCGTCTCCCCCGACTG-3⬘ and 5⬘-CCCAG GGCGTCGAGGAAGAGC-3⬘; for FdmR2, 5⬘-GTGGAGAGAACGACACAT ACGGCGG-3⬘ and 5⬘-GAGGCGGACGACGGGTGGAG-3⬘; for FdmT2, 5⬘-C TCTACAGCTGGACGTTCACCGCG-3⬘ and 5⬘-CCGCTGAGAGGAAGACG GTGGAG-3⬘; for FdmU, 5⬘-GCAACTCCCGCCGCCGAAACC-3⬘ and 5⬘-CA CCCTTGCGGTACGCCGGG-3⬘; for FdmV, 5⬘-CCTGACGAGCAGGGTGT CCGC-3⬘ and 5⬘-GCACGGGCGTACGGGGAGTC-3⬘; and for FdmW, 5⬘-GC GAGGGAGTGGCGGCGC-3⬘ and 5⬘-GAGGCGACGTCACCGTCATGGG3⬘. RT-PCRs were performed with a Qiagen OneStep RT-PCR kit (Qiagen, Valencia, CA) following the manufacturer’s instructions, and 1.0 ␮g of RNA was used for each reaction. The program used was 50°C for 30 min; 95°C for 15 min; 30 cycles of 95°C for 1 min, 60°C for 45 s, and 72°C for 70 s; and a final extension at 72°C for 7 min. Fermentation, isolation, and analysis of FDM A and FDM E. Fermentation, isolation, and analysis of FDM A and FDM E were done essentially by following procedures in the literature (9, 39). Briefly, spores of different strains were inoculated into R2YE with appropriate antibiotics and cultured at 28°C and 250 rpm for 1 or 2 days to prepare seed cultures. Seed cultures were then added to APM (2% [vol/vol]), and the resulting cultures were grown for 8 or 11 days under conditions identical to those used for seed culture production. The fermentation culture was acidified to pH 2.0 with 2 M HCl and centrifuged. The supernatant was discarded, and the pellet (mycelia and any precipitated FDMs) was harvested and extracted three times with acetone. The combined extracts were analyzed by high-performance liquid chromatography (HPLC). HPLC analysis was carried out on a Microsob-MV 100-5 C18 column (5 ␮m by 250 mm by 4.6 mm; Varian, Lake Forest, CA), using a Varian system (Lake Forest, CA). The column was developed with a 20-min linear gradient from 50% CH3CN to 100% CH3CN in H2O with 1% formic acid at a flow rate of 1.0 ml/min. The identities of FDM A and FDM E were confirmed by coinjection

with authentic standards and APCI-mass spectrometry analysis on an Agilent 1100 series LC/MS system (Santa Clara, CA).

RESULTS Bioinformatic analysis of regulatory genes in the fdm cluster. There are three putative regulatory genes in the fdm gene cluster, namely, fdmR, fdmR1, and fdmR2 (39). FdmR1 was proposed to be a SARP-type activator; both FdmR and FdmR2 were found to belong to the MarR family of regulators represented by MarR, the repressor of the multiple antibiotic resistance (mar) operon in E. coli (31, 39). The N-terminal part of FdmR1 (residues 1 to 300) showed significant homology to typical SARP activators, such as DnrI (40.9% identity) from Streptomyces peucetius (35) and ActIIORF4 (33.9% identity) from S. coelicolor (2), featuring both the DNA-binding and transcriptional activation domains (34, 42) and supporting the functional assignment of FdmR1 as a SARP-type activator. The secondary structure of the N-terminal DNA-binding domain of FdmR1 was very similar to that of OmpR, which adopts a winged helix-turn-helix (HTH) moiety (Fig. 2a) (26). The C-terminal part of FdmR1 (residues 301 to 613) showed high similarity with the middle domain of AfsR (residues 301 to 618; 30.5% identity), which is proposed to be an ATPase based on the identified type A and type B ATP binding motifs (Fig. 2b) (34). Notably, a rare TTA codon (nt 574 to 576) for leucine was found in fdmR1. Further indirect evidence supporting FdmR1 as a SARP regulator was revealed by careful examination of the intergenic regions within the fdm cluster; nine putative SARP binding

VOL. 190, 2008


5591

FIG. 3. (a) HPLC analysis of FDM A (}) and FDM E () production in S. griseus. I, ATCC 49344 (8 days); II, SB4013 (8 days); III, SB4014 (8 days); and IV, SB4012 (11 days). (b) Time course of FDM A (}) and FDM E () production in S. griseus ATCC 49344 (open symbols) and SB4012 (filled symbols).

sites containing the specific heptameric direct repeat sequences (TCGAGXX) were found located upstream of the fdmC, fdmD, fdmM, fdmM1, fdmT1, fdmV, fdmR1, fdmW, and fdmR2 genes (Fig. 1b). Each of the postulated SARP binding sites contained the direct repeat sequences strictly spaced by either 4 nt or 15 nt. Spacing between the direct repeats and the ⫺10 regions of each gene was found to be 8 nt, as is typical for promoters regulated by SARPs (Fig. 1c) (34). Inactivation of fdmR1 by gene replacement and complementation of the fdmR1::aac(3)IV mutation by fdmR1 expression in trans. FDM production was completely abolished in S. griseus SB4013 upon inactivation of the fdmR1 gene, and FDM production was restored by expressing fdmR1 in trans. Taken together, these results conclusively established the involvement of FdmR1 in FDM biosynthesis, most likely as an essential activator. The fdmR1 inactivation mutant, S. griseus SB4013, was constructed by replacing a 0.8-kb internal fragment of fdmR1 with aac3(IV), the Am resistance cassette (see Fig. S1a in the supplemental material). The fdmR1::aac(3)IV genotype of SB4013 was confirmed by Southern analysis. Genomic DNAs were digested with BglII and PstI and subjected to Southern analysis, using the 0.48-kb PstI/ClaI internal fragment of fdmR1 as a probe. As expected, the S. griseus wild-type strain revealed one band corresponding to a fragment of 2.4 kb in size; the single-crossover mutant showed two bands, of 2.4 kb and 3.0 kb, while the double-crossover mutant SB4013 gave a band of 3.0 kb (see Fig. S1b in the supplemental material). Subsequently, HPLC analysis revealed that SB4013 failed to produce FDM A and FDM E, thereby supporting FdmR1 as an activator indispensable to FDM biosynthesis (Fig. 3a). Complementation of S. griseus SB4013 was carried out using pBS4047, a middle-copy-number plasmid containing the fdmR1 gene under the control of the strong, constitutive promoter

ermE*. The complemented strain S. griseus SB4014 was constructed by transforming pBS4047 into SB4013, which efficiently restored FDM production, with a titer of approximately twice that of the wild type (Fig. 3a). In the S. griseus ATCC 49344 wild type, the highest yield of FDM (FDM A plus FDM E) was about 242 mg/liter, achieved after 8 days of fermentation in APM. Significantly, the FDM titer was increased to about 506 mg/liter in S. griseus SB4014 cultured under the same conditions (Table 1). Overexpression of fdmR1 in the wild-type S. griseus ATCC 49344 strain. The FDM titer in S. griseus was increased to about 1.36 g/liter by overexpressing fdmR1. Complementation

TABLE 1. FDM titers from all productive strains Strainb

FDM titer (mg/liter) (mean ⫾ SE)a FDM A

FDM E

Total

243 507 1.36 ⫻ 103

S. griseus strains ATCC 49344 SB4014 SB4012c

162 ⫾ 5 235 ⫾ 34 997 ⫾ 30

81 ⫾ 8 272 ⫾ 40 358 ⫾ 29

S. albus strain SB4015

115 ⫾ 6

17 ⫾ 1

S. lividans strains SB4016 SB4017 SB4018 SB4020

ND 0.2 ⫾ 0.04 0.4 ⫾ 0.1 2.2 ⫾ 0.8

ND 0.3 ⫾ 0.1 1.0 ⫾ 0.2 14.5 ⫾ 1.4

132 ND 0.5 1.4 17

a Standard error values are the results of triplicate data acquisition. ND, not detected. b Cultured in APM for 8 days. c Cultured in APM for 10 days.

5592

CHEN ET AL.

of SB4013 with pBS4047 demonstrated that increasing fdmR1 expression can enable S. griseus SB4014 to yield significantly improved titers of FDMs. To further explore the FDM production potential in S. griseus, a pWHM1250-derived highcopy-number plasmid, pBS4045, harboring fdmR1 controlled by the strong, constitutive promoter ermE*, was introduced into the wild-type strain to generate S. griseus SB4012. The FDM production time course of S. griseus SB4012 in APM showed that the FDM titer increased to an impressive level (1,355 mg/liter), about 5.6-fold higher than the yield of the highest yielding wild-type strain, after 10 days of fermentation (Fig. 3a and Table 1). Significantly, the FDM A titer in SB4012 peaked on day 11 (1,036 ⫾ 63 mg/liter), at a level 6.4 times higher than the wild-type titer (162 ⫾ 5 mg/liter on day 8). FDM E, the FDM A precursor, was produced in greatest yield on day 8 (419 ⫾ 61 mg/liter) in SB4012, at a level 4.3-fold greater than the highest FDM E titer in the S. griseus wild-type strain (98 ⫾ 4 mg/liter on day 6) (Fig. 3b). RT-PCR analysis of fdm gene expression in S. griseus wildtype and SB4013 strains. RT-PCR analyses of the transcription of selected genes in S. griseus ATCC 49344 and the fdmR1::aac(3)IV mutant SB4013 revealed that up to 26 of the 28 genes within the fdm cluster, with the exception of fdmR and fdmT2, were positively controlled by FdmR1, directly or indirectly. Intergenic sequence analysis of the fdm gene cluster unveiled nine candidate promoters with SARP binding sites and three other putative promoters. Based on these analyses, 12 genes (fdmC, fdmF, fdmL, fdmO, fdmS, fdmR, fdmU, fdmV, fdmR1, fdmT2, fdmW, and fdmR2) were selected for transcriptional analysis by RT-PCR, each of which was downstream of one putative promoter (Fig. 1b). Total RNAs from the strains were isolated from mycelia grown for 3 days in APM, to the point where FDM production was beginning to increase exponentially. As expected, RT-PCR analysis of fdmR1 showed a strong signal in the S. griseus wild-type strain and no signal in the fdmR1 mutant strain SB4013. Among the other 11 investigated genes, 5 (fdmF, fdmL, fdmO, fdmS, and fdmV) were completely turned off in S. griseus SB4013; all were located in operons downstream of the postulated SARP binding sites. Transcription levels of four genes (fdmC, fdmR2, fdmU, and fdmW) were dramatically reduced in SB4013 relative to those in the wildtype strain. Among the four genes, three (fdmC, fdmR2, and fdmW) were downstream of the putative SARP binding sitecontaining promoters. No such binding site was found in the promoter region upstream of fdmU. The only gene evaluated showing no clear transcriptional difference between the two strains was fdmR, which was found to be controlled by a putative promoter without the typical SARP binding sequence (see Fig. 5a). Notably, in the case of fdmT2, PCR amplification with the appropriate primers afforded the anticipated 1.2-kb fragment. However, no clear signal was observed when fdmT2 was amplified by RT-PCR from both strains, indicating that it could be a pseudogene or that the real gene could be much smaller than the predicted one. In summary, transcripts of all nine genes in operons downstream of promoters containing SARP-type binding sites were shut off or significantly reduced in SB4013. For genes controlled by other putative promoters, fdmU’s transcription level was significantly lower in SB4013, whereas the transcription level of fdmR remained unchanged.

J. BACTERIOL.

FIG. 4. HPLC analysis of FDM A (}) and FDM E () production in fdm heterologous expression strains (at 8 days). I, S. albus SB4015; II, S. lividans SB4016; III, S. lividans SB4017; IV, S. lividans SB4018; V, S. lividans SB4020.

Expression of the fdm gene cluster in heterologous hosts. In heterologous expression experiments, FDMs were produced successfully in S. albus J1074 upon integration of the fdm cluster into the host’s chromosome. However, no FDM production was observed with S. lividans K4-114 as the heterologous host. The whole fdm cluster was inserted into an integrative plasmid, pSET152, to construct pBS4048, which was then introduced into S. albus and S. lividans strains to afford SB4015 and SB4016, respectively. Subsequent fermentation and HPLC analysis of the two strains revealed that S. albus SB4015 produced FDMs with a reasonable titer (132 mg/liter), whereas fermentations of S. lividans SB4016 yielded no FDM (Fig. 4; Table 1). Activation of fdm cluster expression in S. lividans SB4016 by expression of fdmR1 in trans. fdm expression in S. lividans SB4016 was activated, as evidenced by FDM production upon constitutive overexpression of fdmR1 in trans. RT-PCR analysis of the fdm regulatory genes in S. albus SB4015 and S. lividans SB4016 unveiled that transcription of fdmR was slightly lower in SB4015 than that in SB4016; transcription of fdmR2, in contrast, was slightly higher in SB4015. In the case of fdmR1, the gene was transcribed normally in SB4015, in contrast to the complete lack of transcription in SB4016, indicating that the absence of FdmR1 renders S. lividans SB4016 incapable of FDM biosynthesis (Fig. 5b). Increasing the copy number of fdmR1 in S. lividans SB4017, which was realized initially by transforming pBS4049, a highcopy-number plasmid containing fdmR1 with its native promoter, into SB4015, led to a low titer of FDMs (about 0.5 mg/liter). The fdmR1 overexpression strain S. lividans SB4018 was generated by introducing pBS4045, a high-copy-number

VOL. 190, 2008


5593

FIG. 5. (a) RT-PCR analysis of selected fdm genes in the S. griseus ATCC 49344 wild-type and fdmR1-inactivated SB4013 mutant strains. Capital letters indicate the selected genes in the fdm cluster (i.e., C, fdmC; F, fdmF; L, fdmL; O, fdmO; R, fdmR; R1, fdmR1; R2, fdmR2; S, fdmS; U, fdmU; V, fdmV; and W, fdmW). Small letters denote RNAs isolated from the wild-type (w) or mutant (m) strain. (b) RT-PCR analysis of the selected fdm genes in fdm heterologous expression strains. Capital letters indicate the selected genes in the fdm cluster, and small letters denote RNAs isolated from the heterologous expression strains S. albus SB4015 (a), S. lividans SB4016 (b), S. lividans SB4017 (c), and S. lividans SB4018 (d). Equal amounts of RNAs (1 ␮g) were used for the RT-PCRs.

plasmid harboring fdmR1 under the control of the strong, constitutive ermE* promoter, into SB4015. SB4018 produced FDMs with a titer of about 1.4 mg/liter, about three times that of the SB4017 titer, but still very low relative to the 132 mg/liter produced by S. albus SB4015 (Fig. 4; Table 1). RT-PCR analysis of fdm expression in heterologous hosts. Detailed gene transcription analyses of the fdm gene cluster in the four heterologous hosts, S. albus SB4015 and S. lividans SB4016, SB4017, and SB4018, suggested that low expression levels of fdmC might constitute the bottleneck restricting FDM titers in S. lividans strains. Transcription efficiencies of the 12 selected genes were investigated by RT-PCR in the four fdm heterologous expression strains. The transcription levels of fdmR, the only gene not influenced by fdmR1 inactivation, were almost equal in the three S. lividans strains and were slightly higher than that in S. albus SB4015. In the case of fdmR1, transcriptional levels in SB4016, SB4017, and SB4018 were found to be improved incrementally; SB4016 exhibited no fdmR1 transcriptional signal, while both SB4017 and SB4018 showed higher fdmR1 transcription efficiencies than did SB4015. Of SB4017 and SB4018, the latter displayed the greatest fdmR1 transcription. The fdmO and fdmS genes possessed the same transcriptional behaviors as fdmR1 in the four strains. Five other genes, fdmF, fdmL, fdmR2, fdmV, and fdmW, also showed RT-PCR results on par with those for fdmR1. The only difference was that these five genes afforded moderate to weak signals in SB4016. Both fdmC and fdmU displayed no signal in SB4016 and exhibited extremely weak signals in SB4017. Divergently, RT-PCR for fdmC afforded a weak signal in SB4018, whereas fdmU showed a very strong signal (Fig. 5b). Not surprisingly, there was no RT-PCR signal obtained using primers designed for fdmT2 in all four strains. RT-PCR results for all 11 selected genes (except fdmT2) in S. lividans SB4018 revealed that the transcriptional levels for all genes, except for fdmC, were higher than those in S. albus SB4015, indicating that poor transcription of genes in the fdmC-fdmT1 operon could be the cause of the low FDM titers in SB4018 (Fig. 5b). Given the deduced function of FdmC as a putative ketoreductase, a function essential for FDM biosynthesis, expression of fdmC was proposed to be the bottleneck responsible for SB4018’s low FDM titer. Improvement of FDM production by overexpressing fdmC or fdmC in tandem with fdmR1 in S. lividans SB4016. Overexpression of fdmC alone in S. lividans SB4016 did not impact the FDM titer. However, tandem overexpression of fdmC and

fdmR1 in SB4016 afforded FDM titers far superior to those obtained with SB4018. The strain S. lividans SB4019 was generated by transforming pBS4050, an fdmC-overexpressing plasmid, into SB4016; strain SB4020 was similarly derived from SB4016 by introducing pBS4051, containing both fdmR1 and fdmC under the control of ermE*. Following 8 days of fermentation, HPLC analysis revealed SB4019 to be incapable of FDM production. In contrast, SB4020 produced FDMs at a titer of about 17 mg/liter, 12-fold higher than that in SB4018, the SB4016 variant harboring only the FdmR1 overexpression construct (Fig. 4; Table 1). DISCUSSION The fdm gene cluster, responsible for production of the antitumor drug lead FDM A, contains three putative regulatory genes, encoding one SARP-type activator, FdmR1, and two MarR-like regulators, FdmR and FdmR2 (39). FdmR1 was assigned to the SARP class based on the following experimental findings: (i) FdmR1 featured both the N-terminal DNA-binding and transcriptional activation domains common to known SARP regulators; (ii) FDM production was completely abolished in the fdmR1-inactivated S. griseus SB4013 mutant strain, and expressing fdmR1 in SB4013 in trans efficiently restored FDM production; (iii) overexpression of fdmR1 significantly enhanced FDM titers relative to that in the S. griseus wild-type strain; (iv) overexpression of fdmR1 in S. lividans SB4016 afforded this strain the capacity to produce FDMs; and (v) RT-PCR analysis of the S. griseus wild type, the fdmR1-inactivated mutant SB4013, and the fdm heterologous expression strain revealed that up to 26 of the 28 genes within the fdm cluster, with the exception of fdmR and fdmT2, are controlled directly or indirectly by FdmR1, in a positive manner. Incorporation of the ATPase domain along with the SARP N-terminal DNA-binding domain in FdmR1 implies that FdmR1 acts in a fashion similar to that of AfsR, an excellent model for the SARP regulators. This regulator binds to the direct repeat sequences in a dimeric style, recruits RNA polymerase to initiate transcription, and utilizes the energy produced by ATPase activity to accelerate the initiation process (34). Notably, a rare leucine codon, TTA (nt 574 to 576), was found inside the fdmR1 gene. Previous findings with S. coelicolor demonstrated that the tRNA recognizing the UUA codon is specifically encoded by bldA, whose product is accu-

5594

CHEN ET AL.

mulated in late growth phase and plays an important role in developmental regulation, affecting both morphological differentiation and antibiotic production (8). The TTA codon was also discovered in regulatory genes of some other secondary metabolite biosynthetic pathways, such as actII-ORF4 of actinorhodin (13), redZ of undecylprodigiosin (41), and strR of streptomycin (12), which are all biosynthesized during latestage growth. In this study, control of fdmR1 expression by a deduced bldA homolog in S. griseus added an example of how a global regulator (bldA) can control secondary metabolism by turning on a pathway-specific activator (fdmR1). Accordingly, FDMs were not produced until 24 h after inoculation of an S. griseus wild-type seed culture into APM. As a SARP-type activator, FdmR1 positively controls most of the fdm genes. Among them, the operons containing fdmF, fdmL, fdmO, fdmS, and fdmV all possess the SARP binding sites in their promoter regions and were proposed to be controlled primarily by FdmR1. This is consistent with the fact that RT-PCR failed to show any signals for the selected genes in the S. griseus SB4013 fdmR1 mutant. We found that the promoter regions of three other operons, containing fdmC, fdmW, and fdmR2, also had SARP binding sites. Transcripts for fdmC, fdmW, and fdmR2 were found to be reduced in SB4013, indicating that they are at least partially controlled by FdmR1 and may also be activated by other regulators. The proposal that FdmR1 positively regulates the eight operons is also supported by RT-PCR data obtained from the corresponding gene transcripts in the fdm heterologous hosts S. lividans SB4016, SB4017, and SB4018. Transcriptional levels of these genes increased incrementally, coinciding with higher expression levels of fdmR1. Intriguingly, the SARP binding site was also present in the promoter region preceding fdmR1, implying that transcription of fdmR1 is initiated by a global SARP activator or that it may be self-regulated. Among genes downstream of the promoters without SARP binding sites, fdmU was postulated to be influenced by FdmR1 in an indirect manner according to the RT-PCR results; fdmR, in contrast, was completely unaffected by FdmR1. SARP-type activators play a role in many biosynthetic regulatory systems. For instance, DnrI, TylS, and AplV are all involved in pathway-specific regulatory cascades that are tuned by pathway-specific regulators and/or activated by other regulators in the same cascade (1, 3, 32). DnrI is activated by another pathway-specific activator, DnrN, which is positively controlled by DnrO (14). TylS, which is negatively controlled by TylP, can activate the expression of TylR, the activator controlling the tyl biosynthetic genes, with the help of TylU (4). In the case of the fdm biosynthetic gene cluster, FdmR1 directly activates the biosynthetic genes and partially controls a putative repressor gene, fdmR2, whose product might act as a feedback control to avoid accumulation of toxic levels of FDMs. The function of the other postulated repressor, FdmR, has yet to be addressed. Overexpression of defined activator genes has been used to enhance secondary metabolite production (21, 28). This strategy for natural product titer improvement has been applied successfully to some biosynthetic gene clusters bearing SARPtype activators. For example, increased cephamycin and clavulanic acid titers have been achieved by amplification of the ccaR gene in Streptomyces clavuligerus (28), and enhanced

J. BACTERIOL.

mithramycin production has been accomplished via overexpression of mtmR in Streptomyces argillaceus (21). Herein, FDM titers in S. griseus were optimized by expressing fdmR1 under the control of the strong, constitutive ermE* promoter in a high-copy-number plasmid to provide ⬃1.36 g/liter of FDMs. This metabolic engineering approach will vastly hasten the rate at which clinical and basic research studies on FDMs can be conducted. Previous work has shown that when the whole fdm gene cluster is carried on a high-copy-number plasmid and introduced into S. albus J1074, considerable titers of FDMs can be attained (39). When the fdm cluster was cloned into an integrative plasmid and mobilized into both S. albus J1074 and S. lividans K4-114, FDMs were also successfully produced in S. albus SB4015 but not in S. lividans SB4016. The inability of SB4016 to produce FDMs was attributed to the host’s regulatory system defect, since the integrity of the cloned fdm cluster was confirmed by virtue of FDM production in the S. albus host. RT-PCR analysis of the regulatory genes in heterologous hosts revealed that fdmR1 was completely muted in SB4016, indicating that the absence of FdmR1 caused the lack of FDM production in SB4016. Based on this hypothesis, two new S. lividans heterologous strains were constructed to activate fdm expression. One had an increased copy number of fdmR1 under the control of the native promoter (SB4017), and the other had fdmR1 controlled by the strong, constitutive ermE* promoter in a high-copy-number plasmid (SB4018). SB4017 gained the ability to produce FDMs (about 0.5 mg/liter), implying that S. lividans can, in fact, initiate the fdmR1 promoter, albeit with poor efficiency. SB4018 produced more FDMs than SB4017 but still produced FDMs only at low levels (about 1.4 mg/liter). The low FDM titers from these strains prompted us to search for other factors capable of influencing fdm cluster expression in S. lividans. RT-PCR analyses of the four heterologous expression strains revealed that genes in the fdmC-fdmT1 operon might be bottlenecks to FDM biosynthesis in S. lividans K4-114. FdmC, possibly important for polyketide chain assembly, was postulated to be a ketoreductase responsible for the diene moiety formation in FDM A (Fig. 1a). Additionally, FdmT1, a putative transporter, was proposed to serve as a resistance or selfdefense protein (39). Since there were three other genes proposed to encode transporter proteins in the fdm cluster (fdmT, fdmT2, and fdmT3), whose transcriptional modulation might remedy the low FdmT1 levels, the inefficiency of fdmC transcription was postulated to be the main contributor to SB4018’s low FDM titer. To investigate this prospect, the S. lividans strain SB4020 was constructed by overexpressing fdmC and fdmR1 together in SB4016. Fermentation of SB4020 afforded FDMs with a significantly improved titer (about 17.0 mg/liter), indicating that transcription of fdmC, which is controlled cooperatively by FdmR1 and some unidentified regulators, serves an important function in FDM biosynthesis. No FDMs were produced in SB4019, an SB4016 variant that selectively overexpresses only FdmC. These data further substantiate the indispensability of the SARP-type activator FdmR1 in FDM production. Manipulation of the regulatory systems of specific biosynthetic pathways has been proven to be an efficient strategy in achieving heterologous expression of various compounds (15).

VOL. 190, 2008


One example involves expression of the epothilone gene cluster in S. coelicolor CH999 by adding the actI promoter to control the whole gene cluster (36). Similar strategies have been applied successfully in the heterologous expression of 6-methylsalicylic acid in S. coelicolor CH999 (5), macrotetrolides in S. lividans 1326 (18), soraphen A in S. lividans ZX7 (47), and thiocoraline in S. lividans TK24 and S. albus J1074 (22). Another example of heterologous expression efforts capitalizing on acquired knowledge of regulatory systems involves expression of the landomycin biosynthetic gene cluster in a polyketide synthase mutant of Streptomyces fradiae through coexpression of an activator gene, lndI (37). Production of FDMs in S. lividans K4-114, enabled by fdmR1 overexpression, represents another example of this metabolic engineering strategy. Furthermore, it demonstrates that manipulation of not only the pathway-specific regulator but also other important genes, such as fdmC, can significantly influence secondary metabolite titers. Conclusively, FdmR1 was shown to be a SARP-type activator regulating FDM production. RT-PCR studies revealed that up to 26 of the 28 genes within the fdm cluster were controlled by FdmR1. Titers of FDMs were significantly improved, to 1.36 g/liter, by overexpressing fdmR1 in S. griseus ATCC 49344. Moreover, exploitation of fdmR1 overexpression permitted the successful expression of the fdm cluster in heterologous hosts, as exemplified by S. albus J1074 and S. lividans K4-114, opening up the possibility to manipulate FDM biosynthesis in these model Streptomyces species.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

ACKNOWLEDGMENTS We thank the Analytical Instrumentation Center of the School of Pharmacy, University of Wisconsin-Madison, for support in obtaining MS data and Scott Rajski (University of Wisconsin) for assistance in manuscript preparation. This work was supported in part by NIH grants CA78747 and CA113297.

23. 24.

25.

REFERENCES 1. Aigle, B., X. Pang, B. Decaris, and P. Leblond. 2005. Involvement of AlpV, a new member of the Streptomyces antibiotic regulatory protein family, in regulation of the duplicated type II polyketide synthase alp gene cluster in Streptomyces ambofaciens. J. Bacteriol. 187:2490–2500. 2. Arias, P., M. A. Ferna ńdez-Moreno, and F. Malpartida. 1999. Characterization of the pathway-specific positive transcriptional regulator for actinorhodin biosynthesis in Streptomyces coelicolor A3(2) as a DNA-binding protein. J. Bacteriol. 181:6958–6968. 3. Bate, N., G. Stratigopoulos, and E. Cundiffe. 2002. Differential roles of two SARP-encoding regulatory genes during tylosin biosynthesis. Mol. Microbiol. 43:449–458. 4. Bate, N., D. R. D. Bignell, and E. Cundiffe. 2006. Regulation of tylosin biosynthesis involving ‘SARP-helper’ activity. Mol. Microbiol. 62:148–156. 5. Bedford, D. J., E. Schweizer, D. A. Hopwood, and C. Khosla. 1995. Expression of a functional fungal polyketide synthase in the bacterium Streptomyces coelicolor A3(2). J. Bacteriol. 177:4544–4548. 6. Bibb, M. J. 2005. Regulation of secondary metabolism in Streptomyces. Curr. Opin. Microbiol. 8:208–215. 7. Challis, G. L., and D. A. Hopwood. 2003. Synergy and contingency as driving forces for the evolution of multiple secondary metabolite production by Streptomyces species. Proc. Natl. Acad. Sci. USA 100:14555–14561. 8. Chater, K. F. 2006. Streptomyces inside-out: a new perspective on the bacteria that provide us with antibiotics. Philos. Trans. R. Soc. B 361:761–768. 9. Chen, Y., Y. Luo, J. Ju, E. Wendt-Pienkowski, S. R. Rajski, and B. Shen. 2008. Identification of fredericamycin E from Streptomyces griseus: insights into fredericamycin A biosynthesis highlighting carbaspirocycle formation. J. Nat. Prod. 71:431–437. 10. Colombo, V., M. Ferna ńdez-de-Heredia, and F. Malpartida. 2001. A polyketide biosynthetic gene cluster from Streptomyces antibioticus includes a LysR-type transcriptional regulator. Microbiology 147:3082–3092. 11. DeSanti, C. L., and W. R. Strohl. 2003. Characterization of the Streptomyces

26.

27. 28.

29. 30.

31.

32.

33.

34.

35.

5595

sp. strain C5 snp locus and development of snp-derived expression vectors. Appl. Environ. Microbiol. 69:1647–1654. Distler, J., A. Ebert, K. Mansouri, K. Pissowotzki, M. Stockmann, and W. Piepersberg. 1987. Gene cluster for streptomycin biosynthesis in Streptomyces griseus: nucleotide sequence of three genes and analysis of transcriptional activity. Nucleic Acids Res. 15:8041–8056. Ferna ńdez-Moreno, M. A., J. L. Caballero, D. A. Hopwood, and F. Malpartida. 1991. The act cluster contains regulatory and antibiotic export genes, direct targets for transcriptional control by the bldA tRNA gene of Streptomyces. Cell 66:769–780. Furuya, K., and C. R. Hutchinson. 1996. The DnrN protein of Streptomyces peucetius, a pseudo-response regulator, is a DNA-binding protein involved in the regulation of daunorubicin biosynthesis. J. Bacteriol. 178:6310–6318. Galm, U., and B. Shen. 2006. Expression of biosynthetic gene clusters in heterologous hosts for natural product production and combinatorial biosynthesis. Expert Opin. Drug Discov. 1:409–437. Horinouchi, S. 2003. AfsR as an integrator of signals that are sensed by multiple serine/threonine kinases in Streptomyces coelicolor A3(2). J. Ind. Microbiol. Biotechnol. 30:462–467. Kieser, T., M. J. Bibb, M. J. Buttner, K. F. Chater, and D. A. Hopwood. 2000. Practical Streptomyces genetics. The John Innes Foundation, Norwich, United Kingdom. Kwon, H. J., W. C. Smith, L. Xiang, and B. Shen. 2001. Cloning and heterologous expression of the macrotetrolide biosynthetic gene cluster revealed a novel polyketide synthase that lacks an acyl carrier protein. J. Am. Chem. Soc. 123:3385–3386. Latham, M. D., C. K. King, P. Gorycki, T. L. Macdonald, and W. E. Ross. 1989. Inhibition of topoisomerases by fredericamycin A. Cancer Chemother. Pharmacol. 24:167–171. Liou, Y. C., A. Ryo, H. K. Huang, P. J. Lu, R. Bronson, F. Fujimori, T. Uchida, T. Hunter, and K. P. Lu. 2002. Loss of Pin1 function in the mouse causes phenotypes resembling cyclin D1-null phenotypes. Proc. Natl. Acad. Sci. USA 99:1335–1340. Lombo ´, F., A. F. Bran ˜ a, C. Meńdez, and J. A. Salas. 1999. The mithramycin gene cluster of Streptomyces argillaceus contains a positive regulatory gene and two repeated DNA sequences that are located at both ends of the cluster. J. Bacteriol. 181:642–647. Lombo ´, F., A. Velasco, A. Castro, F. de la Calle, A. F. Bran ˜ a, J. M. Sa ńchezPuelles, C. Meńdez, and J. A. Salas. 2006. Deciphering the biosynthesis pathway of the antitumor thiocoraline from a marine actinomycete and its expression in two Streptomyces species. Chembiochem 7:366–376. Lu, K. P., S. D. Hanes, and T. Hunter. 1996. A human peptidyl-prolyl isomerase essential for regulation of mitosis. Nature 380:544–547. MacNeil, D. J., K. M. Gewain, C. L. Ruby, G. Dezeny, P. H. Gibbons, and T. MacNeil. 1992. Analysis of Streptomyces avermitilis genes required for avermectin biosynthesis utilizing a novel integration vector. Gene 111:61–68. Madduri, K., J. Kennedy, G. Rivola, A. Inventi-Solari, S. Filippini, G. Zanuso, A. L. Colombo, K. M. Gewain, J. L. Occi, D. J. MacNeil, and C. R. Hutchinson. 1998. Production of the antitumor drug epirubicin (4⬘-epidoxorubicin) and its precursor by a genetically engineered strain of Streptomyces peucetius. Nat. Biotechnol. 16:69–74. Martıńez-Hackert, E., and A. M. Stock. 1997. The DNA-binding domain of OmpR: crystal structure of a winged helix transcription factor. Structure 5:109–124. Misra, R., and R. C. Pandey. 1982. Fredericamycin A, an antitumor antibiotic of a novel skeletal type. J. Am. Chem. Soc. 104:4478–4479. Pe´rez-Llarena, F. J., P. Liras, A. Rodrı´guez-Garcıá, and J. F. Martıń. 1997. A regulatory gene (ccaR) required for cephamycin and clavulanic acid production in Streptomyces clavuligerus: amplification results in overproduction of both ␤-lactam compounds. J. Bacteriol. 179:2053–2059. Rost, B., G. Yachday, and J. Liu. 2004. The PredictProtein server. Nucleic Acids Res. 32:W321–W326. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Seoane, A. S., and S. B. Levy. 1995. Characterization of MarR, the repressor of the multiple antibiotic resistance (mar) operon in Escherichia coli. J. Bacteriol. 177:3414–3419. Sheldon, P. J., S. B. Busarow, and C. R. Hutchinson. 2002. Mapping the DNA-binding domain and target sequences of the Streptomyces peucetius daunorubicin biosynthesis regulatory protein, DnrI. Mol. Microbiol. 44:449– 460. Sontag, B., J. G. Mu ¨ller, and F. G. Hansske. 2004. Fredericamycin B and C from Streptomyces griseus: structure elucidation after 23 years. J. Antibiot. 57:823–828. Tanaka, A., Y. Takano, Y. Ohnishi, and S. Horinouchi. 2007. AfsR recruits RNA polymerase to the afsS promoter: a model for transcriptional activation by SARPs. J. Mol. Biol. 369:322–333. Tang, L., A. Grimm, Y. X. Zhang, and C. R. Hutchinson. 1996. Purification and characterization of the DNA-binding protein DnrI, a transcriptional factor of daunorubicin biosynthesis in Streptomyces peucetius. Mol. Microbiol. 22:801–813.

5596

CHEN ET AL.

36. Tang, L., S. Shah, L. Chung, J. Carney, L. Katz, C. Khosla, and B. Julien. 2000. Cloning and heterologous expression of the epothilone gene cluster. Science 287:3640–3642. 37. von Mulert, U., A. Luzhetskyy, C. Hofmann, A. Mayer, and A. Bechthold. 2004. Expression of the landomycin biosynthetic gene cluster in a PKS mutant of Streptomyces fradiae is dependent on the coexpression of a putative transcriptional activator gene. FEMS Microbiol. Lett. 230:91–97. 38. Warnick-Pickle, D. J., K. M. Byrne, R. C. Pandey, and R. J. White. 1981. Fredericamycin A, a new antitumor antibiotic. II. Biological properties. J. Antibiot. 34:1402–1407. 39. Wendt-Pienkowski, E., Y. Houng, J. Zhang, B. Li, H. Jiang, H. Kwon, C. R. Hutchinson, and B. Shen. 2005. Cloning, sequencing, analysis, and heterologous expression of the fredericamycin biosynthetic gene cluster from Streptomyces griseus. J. Am. Chem. Soc. 127:16442–16452. 40. Wenzel, S. C., and R. Mu ¨ller. 2005. Recent development towards the heterologous expression of complex of bacterial natural product biosynthetic pathways. Curr. Opin. Biotechnol. 16:594–606. 41. White, J., and M. Bibb. 1997. bldA dependence of undecylprodigiosin pro-

J. BACTERIOL.

42.

43.

44.

45. 46.

47.

duction in Streptomyces coelicolor A3(2) involves a pathway-specific regulatory cascade. J. Bacteriol. 179:627–633. Wietzorrek, A., and M. Bibb. 1997. A novel family of proteins that regulates antibiotic production in streptomycetes appears to contain an OmpR-like DNA-binding fold. Mol. Microbiol. 25:1177–1184. Wilson, D. J., Y. Xue, K. A. Reynolds, and D. H. Sherman. 2001. Characterization and analysis of the PikD regulatory factor in the pikromycin biosynthetic pathway of Streptomyces venezuelae. J. Bacteriol. 183:3468–3475. Yeats, C., S. Bentley, and A. Bateman. 2003. New knowledge from old: in silico discovery of novel protein domains in Streptomyces coelicolor. BMC Microbiol. 3:1–20. Zhang, H., Y. Wang, and B. A. Pfeifer. 2008. Bacterial hosts for natural product production. Mol. Pharm. 5:212–225. Ziermann, R., and M. C. Betlach. 1999. Recombinant polyketide synthesis in Streptomyces: engineering of improved host strains. BioTechniques 26:106– 110. Zirkle, R., J. M. Ligon, and I. Molna ´r. 2004. Heterologous production of the antifungal polyketide antibiotic soraphen A of Sorangium cellulosum So ce26 in Streptomyces lividans. Microbiology 150:2761–2774.