Phage P1-Derived Artificial Chromosomes Facilitate Heterologous Expression of the FK506 Gene Cluster

We describe a procedure for the conjugative transfer of phage P1-derived Artificial Chromosome (PAC) library clones containing large natural product gene clusters (≥70 kilobases) to Streptomyces coelicolor strains that have been engineered for improved heterologous production of natural products. This approach is demonstrated using the gene cluster for FK506 (tacrolimus), a clinically important immunosuppressant of high commercial value. The entire 83.5 kb FK506 gene cluster from Streptomyces tsukubaensis NRRL 18488 present in one 130 kb PAC clone was introduced into four different S. coelicolor derivatives and all produced FK506 and smaller amounts of the related compound FK520. FK506 yields were increased by approximately five-fold (from 1.2 mg L-1 to 5.5 mg L-1) in S. coelicolor M1146 containing the FK506 PAC upon over-expression of the FK506 LuxR regulatory gene fkbN. The PAC-based gene cluster conjugation methodology described here provides a tractable means to evaluate and manipulate FK506 biosynthesis and is readily applicable to other large gene clusters encoding natural products of interest to medicine, agriculture and biotechnology.


Introduction
A wealth of microbial gene clusters that encode the biosynthetic pathways to bioactive natural products, also known as secondary or specialized metabolites, are being unveiled with the everincreasing number of sequenced bacterial genomes [1]. While some clusters correspond to known compounds, a large number (so-called cryptic gene clusters) encode the biosynthesis of previously undiscovered molecules and represent a promising source of new drugs and therapeutics. Many of these gene clusters feature modular type I polyketide synthase (PKS) and/or nonribosomal peptide synthetase (NRPS) enzymes, and can exceed 100 kilobases (kb) in size and consist of several tens of genes [2], [3].
One of the most useful strategies for advancing the study of natural products is heterologous expression of the genes responsible for biosynthesis in a tractable host organism. This approach has many advantages, including compound yield improvements [4], [5], [6] and the ability to work with a strain that is more amenable to genetic manipulation, e. g. in order to disrupt genes [7], to create analogs [8], to study regulation [9], to determine the minimal gene set necessary for biosynthesis [10], or to aid in the identification of unknown molecules, such as from metagenomic libraries [11], [12].
While heterologous expression has been used successfully with small biosynthetic gene clusters (#40 kb) [13], standard techniques are not as straightforward with large biosynthetic gene clusters ($70 kb), largely because of the amount of DNA that needs to be cloned and transferred into a suitable expression host.
In frequently used E. coli cosmid libraries, insert sizes are limited to approximately 42 kb. Thus, large clusters are typically fragmented across multiple clones and require often laborious reassembly strategies such as Red/ET-mediated recombination in E. coli [7], [14] or Transformation Associated Recombination (TAR) in yeast [12], [15].
One method that avoids most problems with cluster reassembly is to clone large gene clusters directly into artificial chromosomes that are able to tolerate insert sizes of approximately 200 kb [16], [17]. This size capacity is sufficient for virtually all known bacterial modular gene clusters and should also allow for several kb of adjacent DNA sequence to be included if the cluster boundary regions are unclear. However, techniques for subsequent handling of these artificial chromosomes and their introduction into heterologous expression strains are not always straightforward. Vectors with large inserts that require protoplast transformation [6], [18] are prone to sequence deletions or rearrangements and frequently give very low transformation rates.
In this paper, we provide an expedient, reliable procedure for transfer of large natural product gene clusters using phage P1derived Artificial Chromosomes (PACs) to Streptomyces coelicolor derivatives that have been engineered for improved heterologous expression. This approach allows for stable integration of PAC DNA via conjugation without any additional modifications, immediately creates a platform for gene cluster analysis, and enables targeted systems and synthetic biology approaches that might not be possible in wild-type strains. To demonstrate the feasibility of this procedure, we have heterologously expressed the 83.5 kb biosynthetic gene cluster for FK506 (tacrolimus), a clinically important polyketide from Streptomyces tsukubaensis NRRL 18488, using a PAC construct of approximately 130 kb.
Both FK506 and the related natural product FK520 (ascomycin) are produced by several Streptomyces strains. The two molecules differ only in one position with the presence of either an allyl (FK506) or ethyl (FK520) side chain ( Figure 1A). Several recent studies have greatly expanded our understanding of FK506 and FK520 biosynthesis, including the identification of the enzyme FkbO as a chorismate hydrolase that generates the first FK506/ FK520 precursor [28], the enzymes responsible for constructing allylmalonyl CoA or ethylmalonyl CoA for the two different side chains of FK506 and FK520 [29], [30], [31] and genes whose products positively or negatively regulate FK506 expression [32], [33]. Introduction of precursors into culture media or overexpression of certain components of the FK506 gene cluster have also led to improvements in FK506 yields [34], [35]. Despite these advances, several challenges exist in pursuing FK506 yield improvements in wild-type Streptomyces producers. Genomic information for FK506 producing Streptomyces has only recently emerged [36], and targeted genetic manipulations in some producing strains are hindered by slow growth or difficulties with introduction of exogenous DNA. In addition, FK506 production is sensitive to a number of culture conditions including pH and the concentrations of lysine, nitrogen, and phosphate [20].
These challenges made FK506 an ideal candidate molecule for heterologous production via PAC based conjugation. Production of FK506 in S. coelicolor should enhance the ability to make targeted gene replacements or modifications and potentially provide new avenues for industrial production.

Results
Generation of a PAC library using the P1-derived artificial chromosome pESAC13 For the generation of a genomic library of S. tsukubaensis NRRL 18488, we used the phage P1-derived artificial chromosome vector ( = PAC vector) pESAC13, depicted in Figure 1B. This vector was developed by M. Sosio and S. Donadio, NAICONS, Milano, Italy and is used by Bio S&T Inc. (Montreal, Canada; see below) for PAC library construction. pESAC13 contains the phage P1 origin of replication, as well as the phage phiC31 integration system that allows stable insertion into the genomes of most Streptomyces species. In contrast to previously developed E. coli-Streptomyces Artificial Chromosomes, pESAC13 (a derivative of pPAC-S1, [16]) contains an oriT site that allows transfer into Streptomyces by conjugation, rather than by time-consuming protoplast transformation. The bla resistance gene in pESAC13 is lost upon library construction; however, the vector also confers kanamycin resistance in E. coli and thiostrepton resistance in Streptomyces.
A major technical challenge in the generation of large insert libraries is the development of protocols for the isolation of highmolecular weight DNA, especially from uncharacterized actinomycetes where even cell lysis can be problematic. This can now be outsourced to companies specializing in this technique; in our case, Bio S&T Inc. (Montreal, Canada; www.biost.com). Starting from  [30]: Yellow, allylmalonyl-CoA biosynthesis; light blue, methoxymalonyl-ACP biosynthesis; dark blue, polyketide synthase; green, chorismate hydrolase (starter unit biosynthesis); brown, peptide synthetase; orange, post-PKS modification; purple, FK506 regulation; red, thioesterase; black, other genes. doi:10.1371/journal.pone.0069319.g001 7 ml flash-frozen packed mycelium, high-molecular weight DNA was isolated by Bio S&T and cloned between the two BamHI sites of pESAC13, which results in the loss of carbenicillin resistance from the parent vector ( Figure 1B). A library of 1920 PAC clones was generated.

Identification of PAC clones containing the complete FK506 gene cluster
Comparison of the previously published sequences of the FK506 biosynthetic gene clusters from different producer strains [30], [36] confirmed that all known FK506 clusters contain a core set of genes including allA-allD, the three PKS genes fkbA-fkbC, and fkbG-fkbQ, with certain strains also containing orthologs of the genes allMNPOS and the regulatory gene fkbR ( Figure 1C). All of these genes are present in an 83.5 kb region of the S. tsukubaensis genome. This region therefore contains all genes known to be involved in FK506 biosynthesis [29], [30] as well as allMNPOS, whose functions have not been elucidated. The recently published genome sequence of S. tsukubaensis NRRL 18488 [36] allowed for bioinformatic investigation of the DNA regions that flank this core cluster. On the left side of the cluster depicted in Figure 1C a putative NRPS siderophore gene cluster was identified that contained putative regulatory genes, while on the right side no genes were identified with any obvious role in secondary metabolite biosynthesis.
The S. tsukubaensis PAC library was screened with the three primer pairs listed in Table 1 and shown in Figure 1C, amplifying fragments flanking and within the cluster. Screening of the first 768 PAC clones ( = two 384 well plates) resulted in two clones showing amplification with all three primer pairs. Gel electrophoresis suggested the inserts to be 160 and 110 kb in size, respectively. Both of these clones were end-sequenced, confirming that they spanned the DNA region containing the FK506 cluster. The smaller PAC, designated 20N, contained the 109347 bp insert depicted in Figure 1C and was chosen for heterologous expression studies.
Integration of the FK506 gene cluster into the genomes of engineered S. coelicolor host strains S. coelicolor is the genetically best-characterized streptomycete, and a large number of tools are available for its genetic manipulation [37], [38], [39]. S. coelicolor M512 [40] has been used previously as a heterologous expression host [13]. Recently, M1146 was developed as an improved host; it lacks all of the genes required for the production of actinorhodin, the prodiginines, coelimycin, methylenomycin and the calcium-dependent antibiotic, and thus lacks anti-microbial activity. The deletion of these biosynthetic gene clusters also removes unwanted sinks of carbon and nitrogen, and hence competition for precursors, and prevents potential cross-talk between different biosynthetic pathways that might result in product inactivation [5]. The strain also has a much simplified extracellular metabolite profile, greatly facilitating analysis of culture supernatants by mass spectrometry. Strains M1152 and M1154 both carry additional mutations in RNA polymerase (in the RpoB subunit) and M1154 also has a mutation in ribosomal protein S12. Both mutations were reported previously to increase the production levels of several secondary metabolites [17], [41], [42].
PAC clone 20N containing the FK506 cluster (PAC20N) as well as a corresponding empty pESAC13 vector (pESAC13Dbla; see Materials and Methods) used as a control were moved into the non-methylating E.coli ET12567 strain [43] via triparental mating and introduced into S. coelicolor M512, M1146, M1152 and M1154 by conjugation, following the procedure summarized in Figure 2.

Heterologous production of FK506
Upon cultivation in MGm 2.5 medium [20], production of FK506 was detected in all strains containing PAC20N using HPLC-UV and HPLC-MS in comparison to an authentic FK506 standard (Inresa Pharma, Bartenheim, France). Figure 3 shows representative chromatograms from strain M1146/PAC20N. Figure S1 shows confirmation of the identity of FK506 by LCMS-MS analysis. No FK506 was detected after the introduction of pESAC13Dbla.
Average production levels in exconjugants from strains M1146, M1152 and M1154 were 0.75, 2.81 and 2.06 mg L 21 , respectively ( Figure 4A), clearly exceeding production in M512 (0.16 mg L 21 ). As observed previously [4], production levels varied among independent exconjugants (e.g. 0.53, 0.68 and 1.05 mg L 21 for three exconjugants of M1146) but were reproducible in repeated cultivations of the same exconjugant strain. A smaller amount of the related compound FK520 was also produced in the strains containing PAC20N, amounting to approximately 10% of FK506 production ( Figure 3). This FK506:FK520 production ratio was comparable to that observed in Streptomyces tsukubaensis NRRL 18488 under the same growth conditions.
Growth rate and production were most reproducible in strain M1146, therefore S. coelicolor M1146/PAC20N exconjugant #3 (1.05 mg L 21 ) was chosen for subsequent experiments to determine if FK506 yields in S. coelicolor could be improved by over-expression of FK506 regulatory genes.
Over-expression of fkbN increases FK506 production levels FK506 production in wild-type Streptomyces strains appears to be positively regulated primarily by the LuxR-type transcriptional regulator FkbN. Over-expression of fkbN increased FK506 yields in S. tsukubaensis NRRL 18488 [33] and Streptomyces sp. KCTC 11604BP [32] and fkbN is present in all other identified FK506 gene clusters [30]. We introduced fkbN into S. coelicolor M1146/ PAC20N#3 using the replicative, conjugative expression plasmid pUWL-oriT-apr containing the strong constitutive ermE* promoter [44] and monitored FK506 production levels as above. While growth rates and biomass accumulation were comparable, overexpression of fkbN resulted in approximately five-fold FK506 yield improvements (5.5 mg L 21 compared to 1.2 mg L 21 observed upon introduction of an empty pUWL-oriT-apr plasmid; Figure 4B), demonstrating that fkbN over-expression is also beneficial to FK506 yields when the cluster is expressed heterologously.
The LysR-type regulator FkbR has also been shown to influence FK506 production levels, although this gene or its homologues have not been observed in all FK506 gene clusters [30]. In Streptomyces sp. KCTC 11604BP, over-expression of the fkbR homologue tcs7 decreased FK506 production, while tcs7 deletion elevated FK506 yields [32]. In contrast, in S. tsukubaensis NRRL 18488 over-expression of fkbR increased FK506 production levels, while deletion had the opposite effect [33]. We introduced fkbR into S. coelicolor M1146/PAC20N#3 in the same manner as fkbN, but observed a decrease of 44% (0.69 mg L 21 , Figure 4B). It appears that constitutive, high fkbR expression in S. coelicolor is detrimental to FK506 yields.
We also generated a pUWL-oriT-apr expression plasmid containing tandemly arranged fkbN and fkbR (each with the same ribosome binding site that was used for their individual expression) that resulted in an increase of FK506 production (2.86 mg L 21 ), but clearly less than the highest yields obtained when overexpressing fkbN alone ( Figure 4B). This is consistent with the positive regulatory effect of fkbN over-expression being counteracted by a negative effect of constitutive expression of fkbR.

Discussion
FK506 (tacrolimus) is the most important immunosuppressant in clinical use [45]. Substantial effort has led to yield improvements in wild-type producers through regulator manipulation [32], [33], precursor supplementation into culture media [31], [34], [46] or by introducing extra copies of FK506 biosynthetic enzymes [35]. To the best of our knowledge, our PAC-based conjugation strategy has afforded the first example of heterologous FK506 production in a more tractable expression host. Transfer of the FK506 gene cluster to S. coelicolor removes any obstacles inherent in working with native producer strains, including issues with sporulation, suitable growth media, genetic instability, or genetic manipulation. Additionally, while genetic modifications in wild-type strains require time-consuming double crossover experiments, PAC introduction into heterologous strains such as S. coelicolor permits rapid modification of the gene cluster in E. coli and direct examination of the primary exconjugants in the heterologous host, thereby reducing the amount of steps needed for such investigations. It also allows use of the wide variety of genetic tools available for S. coelicolor. Production of FK506 in S. coelicolor confirms that all genes necessary for its synthesis in a heterologous streptomycete host are present in our PAC clone, which simplifies future analysis of FK506 and precursor biosynthesis.
Although initial FK506 yields in S. coelicolor M1146 were modest (up to approximately 1 mg L 21 in microtiter plates, Figure 4) significant yield improvements were observed upon over-expression of the positive FK506 regulator fkbN (5.5 mg L 21 ) and we anticipate additional modifications should continue to increase heterologous production yields. However, FK506 production decreased upon over-expression of fkbR in S. coelicolor M1146/ PAC20N#3, which is contrary to recent evidence showing the positive effect of fkbR on FK506 production levels in S. tsukubaensis NRRL 18488 [33]. This result may indicate that fkbR functions differently in S. coelicolor, perhaps by affecting an unknown component of global regulation that has some influence on FK506 biosynthesis. It is also possible that high level expression of fkbR is detrimental to FK506 biosynthesis, and that this regulator might normally be transcribed at lower levels and/or in a timecontrolled fashion. This finding may be reflected in the opposing conclusions reported in other studies of fkbR [32], [33].
The triparental mating and conjugation procedure we describe here for stable transfer of PAC DNA should be applicable to many other natural product gene clusters of biomedical and/or biotechnological interest. Actinobacteria are among the most prolific microbial producers of natural products and the majority of known antibiotics are derived from them [47], [48]. Several antibiotics from actinobacteria are encoded by large gene clusters such as enduracidin (84 kb, [49]), and rubradirin (105 kb, [2]), and could also be useful targets for heterologous expression via PAC cloning. With increasing amounts of data available from genome sequencing projects, PAC-based strategies hold significant value and promise in simplifying subsequent biosynthetic investigations. These methods may be especially valuable when large unidentified gene clusters appear to be ''silent'', only produce trace amounts of compound that complicate or prohibit structural elucidation, or are found in rare or intractable bacterial strains that do not have any genetic techniques established. In addition, PAC-based conjugation could serve as a means of mobilizing large artificially derived natural product gene clusters as gene synthesis technologies become more affordable. Future work with the PAC containing the FK506 gene cluster or with other PACs containing gene clusters of high relevance could also be targeted at improving industrial compound yields by testing a variety of heterologous hosts and eventually lowering production costs. We envision that large insert DNA libraries will become a standard complement to microbial genome sequencing projects in the near future [6]. The amount of available sequence information outpaces experimental verification of gene function, and although the quality of bioinformatics-based prediction of natural product gene clusters has advanced tremendously over the last decade [50] strategies to accelerate gene cluster identification and structure elucidation and increase compound production are urgently needed. The PAC based methodology described here to mobilize natural product gene clusters of virtually any size to engineered heterologous Streptomyces production strains should contribute to shortening this timeline and remove impediments to genetic or system biology studies aimed at improving access to compounds of interest.

Preparation of Phage P1 Artificial Chromosome (PAC) library
Streptomyces tsukubaensis NRRL 18488 was obtained from the US Department of Agriculture strain collection and grown on ISP4 plates [20]. Mycelia were prepared for PAC library construction by growing S. tsukubaensis in YEME medium (100 ml) inoculated with mycelia from a TSB starter culture (10 ml, [37]). The mycelia were grown in YEME for 72 h at 30uC in a 300 ml baffled flask with a stainless steel spring, and were then centrifuged to remove the media. The mycelial pellets were washed three times (200 mM NaCl, 10 mM Tris-Cl (pH 7.2), 100 mM EDTA (pH 8.0)) and flash frozen in liquid nitrogen before being stored at 280uC. PCR primers (Table 1) were designed for PAC library screening to isolate positive clones containing the entire FK506 gene cluster, and were tested beforehand using S. tsukubaensis genomic DNA as PCR template. Each PCR product was sequenced to verify that the intended regions of the gene cluster were amplified. Figure 2. Procedure for introduction of PAC DNA into Streptomyces coelicolor. A) Two DH10b strains, containing either the pR9406 helper plasmid (carbenicillin resistance) or the FK506 PAC (kanamycin resistance in E. coli, thiostrepton resistance in S. coelicolor), are triparentally mated on LB agar without antibiotics with the non-methylating damdcm -ET12567 strain (chloramphenicol resistance). Following the mating, the resultant E. coli patch is restreaked to select for ET12567 cells that contain both vectors. B) Resistant colonies are grown to an OD 600 of 0.4, conjugated with S. coelicolor spores and plated on R2 medium without sucrose [37]. C) Additional plating and inoculation steps to prepare S. coelicolor for production medium. thio = thiostrepton, ndx = nalidixic acid. doi:10.1371/journal.pone.0069319.g002 Approximately 7 ml of concentrated mycelia were used to construct a phage P-1 artificial chromosome library using a modified E. coli -Streptomyces Artificial Chromosome (ESAC, [16]) vector (pESAC13, Figure 1B). This vector contains an oriT, a phiC31 integrase gene and a phiC31 attP site that permit conjugation into Streptomyces strains and integration at the attB recombination locus.
The pESAC13 library construction was performed by Bio S&T (Montreal, Canada). Vector DNA was digested with BamHI, dephosphorylated and purified using standard procedures. Highmolecular-weight (HMW) DNA was isolated by embedding the mycelium in 1% low-melting-point agarose plugs that were treated with proteinase K overnight, and then partially-digested with Sau3AI. The partially digested HMW DNA was size-fractionated on a 1% (w/v) pulsed field agarose gel in 0.5X TBE using a CHEF DRIII (Bio-Rad, Canada). The 100-250 kb DNA fragments were eluted from the gel by PFGE and dialyzed against 1X TE (10 mM Tris-HCl, 1 mM EDTA, (pH 8.0)) prior to ligation with the vector. The ligation mix was used to transform 20 ml of E. coli DH10b cells (Invitrogen, USA) by electroporation using a CellPorator equipped with a voltage booster (Invitrogen, USA). Transformants were selected at 37uC on LB medium supplemented with 20 mg ml 21 kanamycin and 5% sucrose. For quality control, insert size was determined by DraI or NotI digestion of mini-prepped PAC DNA and subsequent PFGE gel separation. The average insert size of the pESAC library was estimated to be 125 kb.
Two PACs (approximated by gel electrophoresis to contain inserts of 160 and 110 kb) were identified by PCR screening as likely to contain the entire FK506 gene cluster. Both PACs were end sequenced, and the smaller of the two PACS (PAC20N, determined by end sequencing to contain a 109347 bp insert) was chosen for heterologous expression ( Figure 1C). The DH10b E. coli strain containing PAC20N was used in a triparental mating ( Figure 2) with another DH10b E. coli strain that contained the driver plasmid pR9406 (derived from pUB307 [51]), in which the kanamycin marker has been replaced with carbenicillin) and the non-methylating E. coli strain ET12567. Using overnight starter cultures, these three strains were grown to an OD 600 of 0.4 and mixed together on LB plates without antibiotic selection. After 24 h, the resulting E. coli patch was streaked on selective (kan R , carb R , chl R ) LB plates. The resulting ET12567 exconjugants were tested by PCR to confirm that they contained the entire PAC20N clone insert. To create a negative control strain to use in parallel to the PAC20N clone, pESAC13 was digested with BamHI and ScaI and religated to remove the bla resistance gene. This empty vector, pESAC13Dbla, was also transferred to ET12567 by triparental mating for later conjugation into S. coelicolor.

Introduction of PACs into S. coelicolor strains
The ET12567 strains containing pR9406 and either PAC20N or pESAC13Dbla were conjugated with germlings of S. coelicolor M512, M1146, M1152, and M1154 [5] according to standard protocols [37] (Figure 2). Following conjugation, mixtures of spores and E. coli were plated on R2 medium without sucrose and overlaid after 16-20 h with thiostrepton (60 mg ml 21 ) and nalidixic acid (25 mg ml 21 ). Exconjugants were streaked on DNA plates containing thiostrepton and nalidixic acid at the same concentrations, and then on MS (SFM) plates with 10 mM MgCl 2 and nalidixic acid to allow colonies to sporulate. Three sporulating colonies from each S. coelicolor strain were spread on new MS+MgCl 2 + nalidixic acid plates and allowed to grow for 5- 7 d prior to harvesting spores for inoculating cultures. After harvesting, dilution series of spores were spread on MS plates to obtain concentrations for each spore stock.
Spores (5610 8 ) of each strain were inoculated into 50 ml 1:1 TSB:YEME cultures containing thiostrepton (10 mg ml 21 ). Each culture was grown for 3 d at 30uC in baffled flasks as described above. The mycelia were collected and frozen overnight in 20% peptone (7.5 ml total volume in 50 ml Falcon tubes). The mycelia were subsequently homogenized [52] and stored at 280uC to serve as stocks for inoculation of production medium. To measure heterologous production of FK506, each strain was grown in 24well microtiter plates (28u, 300 RPM, [4,52]) for 6 days in MGm 2.5 medium ( [20] with minor modification). Each well of the 24well microtiter plate contained 3 ml of MGm 2.5 media and was inoculated with 75 ml of homogenized mycelia (3 wells per culture replicate). For quantification of FK506 production in each S. coelicolor strain, PAC20N and pESAC13Dbla integrants were grown in triplicate.

Detection of FK506 and FK520
Culture samples from each S. coelicolor strain were extracted with ethyl acetate (1:1) and each extract was reduced to dryness using a speed vacuum system. Samples were resuspended in 100 ml of methanol and profiled using HPLC (25 ml injections) and LCMS (2.5 ml injections) in comparison to standards of FK506 (Inresa Pharma, Bartenheim, France) and FK520 (Sigma). An Genetic modifications to S. coelicolor M1146 containing the FK506 PAC The genes encoding the FK506 regulators FkbN and FkbR were PCR amplified from the FK506 PAC and ligated independently into separate pUWL-oriT-apr vectors containing the ermE* promoter [44], [53]. The forward primer for each amplicon (Table 1) contained a Streptomyces ribosome binding site [54]. These constructs were verified by sequencing and conjugated, along with pUWL-oriT-apr as a control, into S. coelicolor M1146/PAC20N exconjugant #3 using standard protocols [37], [43]. This S. coelicolor M1146 FK506 derivative had shown the highest levels of FK506 production when assaying S. coelicolor M1146/PAC20N exconjugants in preliminary experiments. Another pUWL-oriT-apr construct was created containing both fkbN and fkbR, with each gene preceded by the same ribosome binding site as above [54]. All plates, preculture media and production media used for  [52]. Data show mean values and standard errors of all measurements. (B) Effect of over-expression of the regulator genes fkbN and fkbR on FK506 production in S. coelicolor M1146/PAC20N#3 as compared to M1146/PAC20N#3 containing an empty pUWL-oriT-apr vector. Averages from the highest producing exconjugants from assays testing M1146/PAC20N#3 + fkbN, + fkbR and fkbN + fkbR are shown. doi:10.1371/journal.pone.0069319.g004 growing S. coelicolor strains containing the pUWL-oriT-apr constructs were maintained under apramycin selection. Figure S1 LCMS-MS analysis of FK506 standard (top) and FK506 produced by Streptomyces coelicolor M1146/ PAC20N#3 (bottom). The parent ion is the sodium adduct of FK506 (826.7 m/z). (TIF)