Functional characterization and manipulation of ... - Wiley Online Library

8 downloads 49046 Views 441KB Size Report
Feb 22, 2010 - advances in fungal gene manipulation tools present vast opportunities for ..... (APS123, APS15 and FAS4) was performed by an automated.
Molecular Microbiology (2010) 76(2), 456–466 䊏

doi:10.1111/j.1365-2958.2010.07109.x First published online 31 March 2010

Functional characterization and manipulation of the apicidin biosynthetic pathway in Fusarium semitectum mmi_7109 456..466

Jian-Ming Jin,1† Seunghoon Lee,1 Jungkwan Lee,1 Seung-Ryul Baek,1‡ Jin-Cheol Kim,2 Sung-Hwan Yun,3 Sook-Young Park,4§ Seogchan Kang4 and Yin-Won Lee1* 1 Department of Agricultural Biotechnology and Centers for Fungal Pathogenesis and for Agricultural Biomaterials, Seoul National University, Seoul 151-921, Korea. 2 Chemical Biotechnology Center, Korea Research Institute of Chemical Technology, Daejon 305-600, Korea. 3 Department of Medical Biotechnology, Soonchunhyang University, Asan 336-745, Korea. 4 Department of Plant Pathology, The Pennsylvania State University, University Park, PA 16802, USA.

Summary Apicidin is a cyclic tetrapeptide produced by certain isolates of Fusarium semitectum and has been shown to inhibit Apicomplexan histone deacetylase. An apicidin-producing strain (KCTC16676) of the filamentous fungus was mutated using an Agrobacterium tumefaciens-mediated transformation, resulting in 24 apicidin-deficient mutants. Three of the mutants had a T-DNA insertion in a gene that encodes a nonribosomal peptide synthetase (NRPS). Results of sequence, expression, and gene deletion analyses defined an apicidin biosynthetic gene cluster, and the NRPS gene was named as apicidin synthetase gene 1 (APS1). A 63 kb region surrounding APS1 was sequenced and analysis revealed the presence of 19 genes. All of the genes including APS1 were individually deleted to determine their roles in apicidin biosynthesis. Chemical analyses of the mutant strains showed that eight genes are required for apicidin production and were used to propose an apicidin biosynthetic pathway. The apicidin analogues apicidin E, apicidin D2 and apicidin B were identified Accepted 22 February, 2010. *For correspondence. E-mail lee2443@ snu.ac.kr; Tel. (+82) 2880 4671; Fax (+82) 2873 2317. Present addresses: †Department of Biology, Hong Kong Baptist University, Kowloon, Hong Kong, China; ‡Daewon Pharm. Ltd, Seoul 143-150, Korea; §Center for Fungal Pathogenesis, Seoul National University, Seoul 151-921, Korea.

© 2010 Blackwell Publishing Ltd

from chemical analysis of the mutants. The cluster gene APS2, a putative transcription factor, was shown to regulate expression of the genes in the cluster and overexpression of APS2 increased apicidin production. This study establishes the apicidin biosynthetic pathway and provides new opportunities to improve the production of apicidin and produce new analogues.

Introduction Considering the diverse metabolic capabilities in the limited number of commercially utilized fungal species and the under-explored fungal diversity, fungi represent a vastly under-utilized resource for new pharmaceuticals and other functional compounds. Rapid accumulation of sequenced genomes from diverse fungal species and advances in fungal gene manipulation tools present vast opportunities for uncovering and manipulating novel metabolic capabilities for the production of useful compounds. The genus Fusarium includes many species that produce a plethora of secondary metabolites, some of which are toxic to animals and humans (Desjardins, 2006). Certain isolates of F. semitectum produce cyclic tetrapeptides [cyclo-(L-N-O-methyl-tryptophanylL-isoleucinyl-D-pipecolinyl-L-2-amino-8-oxo-decanoyl)] named apicidin (1), apicidin D2 (3) and apicidin B (4) (Fig. 1) (Singh et al., 1996; 2001; 2002; Park et al., 1999; Jin et al., 2008). Apicidin exhibits broad-spectrum antiprotozoal activity in vivo against Apicomplexan parasites. Antiprotozoal activity of apicidin results from its inhibition of Apicomplexan histone deacetylase, a key nuclear enzyme involved in transcriptional control (Darkin-Rattray et al., 1996). In addition, apicidin inhibits the growth of cancer cells and human endometrial cells (Darkin-Rattray et al., 1996; Han et al., 2000; Hong et al., 2003; Ueda et al., 2007). The cyclic nature of the peptide, the C-8 keto group, and the tryptophan residue are critical for its biological activity (Singh et al., 1996). Cyclic peptides are synthesized by non-ribosomal peptide synthetases (NRPSs) in bacteria and fungi (Doekel and Marahiel, 2001; Finking and Marahiel, 2004). NRPSs consist of several multifunctional enzymatic modules. Each module contains three core domains that perform condensation, adenylation and thiolation, with the

Apicidin biosynthetic genes in Fusarium semitectum 457

Fig. 1. Structures of apicidin (1), apicidin E (2), apicidin D2 (3) and apicidin B (4).

first module usually lacking a condensation domain. In addition to these three core domains, some modules may contain additional domains, which perform epimerization, N-methylation or heterocyclization. Such modifications are important for biological activity of the resulting cyclic peptides. Each module incorporates a specific amino acid, which is subsequently modified by the extra domains, and thus the number and order of the modules within each NRPS dictates the sequence of the resulting cyclic peptide (Doekel and Marahiel, 2001; Finking and Marahiel, 2004). Fungi often use a single NRPS gene to synthesize the whole peptide backbone; whereas in bacteria, more than one NRPS gene participates in the synthesis of a single peptide. In fungi, HC-toxin of Cochliobolus carbonum (Scott-Craig et al., 1992), enniatin of F. scirpi (Haese et al., 1993) and cyclosporin of Tolypocladium niveum (Weber et al., 1994) are examples of cyclic peptides produced by a single NRPS gene. The unique structure, mode of action and biological activity of apicidin have generated strong interest in the possibility of producing novel apicidin analogues by isolating strains with such capability and/or genetic modifications of the existing biosynthetic pathway (Singh et al., 1996). In this study, we cloned and defined the genes involved in apicidin biosynthesis from an apicidin-producing F. semitectum strain (Park et al., 1999; Jin et al., 2008). Reconstruction of the biosynthetic pathway via targeted mutagenesis of individual genes and chemical analyses of the resulting mutants allowed the engineering of this biosynthetic pathway to improve the production of apicidin and to produce new analogues.

Results Cloning of the apicidin biosynthetic genes Random insertional mutants of F. semitectum KCTC16676 were generated by Agrobacterium tumefaciens-mediated transformation (ATMT). Co-cultivation of F. semitectum

mycelia with A. tumefaciens cells in the presence of acetosyringone (AS) led to the generation of hygromycin B-resistant transformants approximately 10 days after transfer to selective medium amended with hygromycin B. An assessment of the mitotic stability of 10 randomly selected transformants showed that all maintained their hygromycin B-resistant phenotype after being subcultured five times in the absence of hygromycin B, suggesting that the inserted selection marker is mitotically stable. The optimized conditions for ATMT in F. semitectum KCTC16676 consisted of a 6 h pretreatment of A. tumefaciens cells in the induction medium (IM) containing AS and the co-cultivation of fungal conidia (1 ¥ 106 ml-1) with an equal volume of bacterial cells for 48 h. We generated 2325 transformants of which apicidin production ability was screened by thin layer chromatography (TLC). Twenty-four transformants failed to produce apicidin, but their growth rate, pigmentation and conidiation were indistinguishable from the wild-type strain. The lack of apicidin production in the 24 transformants was confirmed by high-pressure liquid chromatography (HPLC) analysis. Southern analysis showed that 16 mutants contained a single copy of T-DNA. We employed thermal asymmetric interlaced polymerase chain reaction (TAIL-PCR) to isolate genomic DNA segments adjacent to inserted T-DNA in the 16 mutants. Sequences of three of the resulting amplicons showed an insertion of T-DNA in a gene with high similarity to the C. carbonum gene HTS, which encodes an NRPS required for HC-toxin synthesis (M98024, 43% identity; Scott-Craig et al., 1992). We tentatively designated this gene as apicidin synthetase 1 (APS1). Approximately 5000 cosmid clones of KCTC16676 were screened with a 559 bp PCR product corresponding to APS1 as a probe to clone the APS1 and neighbouring genes, resulting in two positive clones (APS123 and APS15). A probe based on sequence near the end of APS15 was used to isolate a third cosmid clone, FAS4, which overlapped APS15 (Fig. 2). These cosmids were

© 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466

458 J.-M. Jin et al. 䊏

Fig. 2. Organization of the genes in the apicidin biosynthesis gene cluster and schematic diagram of the modules in the apicidin synthetase encoded by APS1. Sequencing of three overlapping cosmid clones covered a 63 kb genomic region containing 19 ORFs. The arrows denote predicted genes and their transcriptional direction; black arrows indicate the genes required for apicidin biosynthesis and white arrows indicate the genes that are not necessary for apicidin production. The function of APS11 (gray arrow) could not be confirmed because we failed to delete APS11. The vertical arrow under APS1 indicates the T-DNA insertion site.

subjected to shotgun sequencing, and any remaining gaps were filled by targeted subcloning and primer walking, yielding ~63 kb of contiguous DNA sequence (GenBank Accession No. GQ331953) that contains 19 putative open reading frames (ORFs) (Fig. 2). Functional analyses of the genes in the APS gene cluster To determine which of the 19 genes located in the ~63 kb region might be involved in apicidin production, we com-

pared their expression profiles by RT-PCR under apicidin producing (NPF2 medium) and non-producing (complete medium) conditions. In this region, the 10 genes immediately upstream of APS1 exhibited patterns of expression that were similar to the expression pattern of APS1, and those genes were named as APS2-APS11 (Fig. 2). In contrast, genes downstream or further upstream of APS1 exhibited different patterns of expression. Putative functions of the proteins encoded by these genes were assigned based on comparison to known genes in the databases by BLAST analysis (Table 1).

Table 1. Predicted functionsa of genes located in and adjacent to the apicidin biosynthesis gene cluster. Most similar/informative homologue Gene

Size (bp)b

Predicted function

Accession No.

Identity (%)

APS1 APS2 APS3 APS4 APS5 APS6 APS7 APS8 APS9 APS10 APS11 ORF-1 ORF-2 ORF-3 ORF-4 ORF-5 ORF-6 ORF-7 ORF-8

15 432 1 146 918 1 104 4 857 1 173 1 596 1 485 1 725 753 1 710 1 101 2 595 528 2 853 1 743 2 355 1 944 1 419

Non-ribosomal peptide synthetase Transcription factor Pyrroline reductase Aminotransferase Fatty acid synthase O-methyltransferase Cytochrome P450 monooxygenase Cytochrome P450 monooxygenase Oxidase Reductase Efflux carrier Racemase MFS transporter Esterase/lipase Hypothetical protein Hypothetical protein Glucuronyl hydrolase Transcription factor Transcription factor

Q01886 (HC-toxin synthetase HTS) O74205 (Transcription factor TOXE) NP_821060 Q9Y885 (Aminotransferase TOXF) 2UV8_A ABE60721 XP_001275091 XP_001275091 XP_754868 AAS90000 Q00357 (Putative HC-toxin efflux carrier TOXA) ZP_02736918 XP_002378547 XP_747967 EEU35424 XP_382723 XP_749875 CBF86816 CBF86816

43 31 32 56 44 27 46 33 36 39 46 40 39 61 53 39 17 33 28

a. Functions were predicted based on similarity of the genes to genes with known functions. b. The length of the coding region.

© 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466

Apicidin biosynthetic genes in Fusarium semitectum 459

Fig. 3. Disruption of APS1 via homologous recombination. A. Schematic diagram illustrating the gene replacement strategy. The probe used for hybridization was amplified from wild-type strain KCTC16676. B. Southern hybridization of genomic DNAs digested with XhoI. Lane 1, wild-type strain KCTC16676; lanes 2-4, APS1-disrupted mutants; lane 5, an ectopic transformant. APS1-disrupted mutants are expected to carry a single 5.1 kb band instead of the 3.7 kb band present in KCTC16676.

The APS1 gene has a 15 488 bp ORF with a single 56 bp putative intron and is predicted to encode a protein (567 kDa) that consists of four peptide synthase modules with an epimerization domain in the first module (Fig. 2). To confirm that APS1 is essential for apicidin production, a transformation-mediated approach was used to replace a 400 bp fragment, corresponding to a 10 990–11 390 bp region downstream from the APS1 translational start site in F. semitectum KCTC16676 with a hygromycin resistance gene (hygB) cassette. Of 18 transformants, three disrupted the gene (Fig. 3) and failed to produce apicidin; by contrast the wild-type strain and transformants without the disruption produced apicidin and its analogues (Fig. 4A and B). To determine whether the genes co-regulated with APS1 (APS2-APS11) or the non-co-regulated genes (ORFs 1–3 and ORFs 5–7) flanking this gene cluster are also required for apicidin biosynthesis, we deleted them individually via targeted mutagenesis, which was confirmed by Southern analyses (Fig. S1). Mutants aps2, aps5 and aps6 produced neither apicidin nor any known apicidin analogues. Deletion of APS4, APS10, and the six genes neighbouring the APS cluster had no effect on apicidin production. Repeated attempts to delete APS11, encoding a protein similar to a putative HC-toxin efflux carrier in C. carbonum (Pitkin et al., 1996), were unsuccessful. Mutants aps3 and aps9 produced apicidin B and apicidin D2, respectively, instead of apicidin (Fig. 4C and D). These metabolites were confirmed by LC-MS and NMR (Table S1). Identification of a novel apicidin analogue Mutant aps7 produced no apicidin, but produced a novel apicidin analogue designated as ‘2’ (Figs 1 and 4E). The

molecular formula of this compound was determined to be C34H51N5O5 by 13C DEPT NMR and electrospray ionization mass spectrometry (EIMS), which showed a molecular ion at m/z 609. Its 1H and 13C NMR spectra were identical to those of apicidin, including NH shifts, except that the signal for the C-8 ketone group was absent in the 13C NMR spectra and the signal for a methylene group at dC 29.29 was present. In the HMC spectrum of ‘2’, the proton signal at dH 0.88 (3H, dd) was correlated with carbon signals at dc 29.29 (CH2) and 22.62 (CH2), which were assigned as C-8 and C-9 of the new amino acid residue. Thus, its structure was determined to be [cyclo-(L-N-O-methl-tryptophanyl-Lisoleucinyl-D-pipecolinyl-L-(2-amino-decanoyl)], and this new apicidin analogue was named as apicidin E (Table S1, Fig. 1). In addition, aps8 mutant failed to produce apicidin or the other analogues, but accumu-

Fig. 4. HPLC analysis of apicidin and its analogues. A. Wild-type strain. B. APS1-disrupted mutant. C. APS3-deleted mutant. D. APS9-deleted mutant. E. APS7-deleted mutant. F. APS8-deleted mutant. Samples shown in (A–D) and (F) were eluted with 75% methanol and the sample in (E) was eluted with 85% methanol. The numbers above the peaks indicate the following: 1, apicidin; 2, apicidin E; 3, apicidin D2; 4, apicidin B; and 5, unknown compound.

© 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466

460 J.-M. Jin et al. 䊏

APS11 were transcribed in both aps2 mutant and the wild-type strain at similar levels at day 2, but their transcripts increased only in the wild-type strain as time progressed. APS3 transcripts were present in both aps2 mutant and the wild-type strain, but its level increased only in the wild-type strain. Thus, expression of the genes in the APS gene cluster is under APS2 regulation. Transcription patterns of other genes, including APS1 and APS5–APS10, were similar to that of APS2 (Fig. 6). ORF1 was constitutively expressed regardless of the presence of APS2, and ORF5 was not expressed in all strains.

Fig. 5. Time-course of apicidin production. Apicidin was extracted from cultures of the wild-type strain KCTC16676 (WT) and APS2 overexpression mutant (APS2-OX1) grown in NPF2.

lated a new, structurally related material designated as ‘5’ (Fig. 4F). APS2 regulates the expression of genes in the APS gene cluster APS2 encodes a putative transcription factor (Table 1), and loss of its function abolished apicidin production. Complementation of the aps2 mutant by introducing a wild-type copy of the gene fully restored apicidin production, confirming that APS2 is required for apicidin production. Increased expression of APS2 using the promoter of the b-tubulin gene of F. graminearum caused earlier and higher production of apicidin compared with the wild-type strain 2, 4, 6 and 8 days after culturing in NPF2 medium at 25°C (Fig. 5). Three APS2overexpressing transformants (APS2-OX1 to APS2-OX3) produced five times more apicidin than the wild-type strain in rice cultures. Expression patterns of the genes in the cluster in aps2 mutant, wild-type strain and APS2-OX1 mutant were compared by northern analysis. APS2 transcripts were first detected in 6-day-old cultures of the wild-type strain but not in aps2 mutant (Fig. 6). APS4 and

Survey of Fusarium species for the ability to produce apicidin We previously screened 23 strains in the F. semitectum/F. equiseti species complex for apicidin production, none of which produced apicidin (Park et al., 1999). Thirty-eight Fusarium isolates representing the F. graminearum species complex and closely related species (Table S2) were screened first for the presence of APS genes. Three F. graminearum lineage 6 (F. asiaticum) isolates exhibited hybridization signals to probes APS7, APS11 and APS1 (Fig. S2), but none of them produced apicidin.

Discussion Agrobacterium tumefaciens-mediated transformation led to the isolation of the genes required for apicidin biosynthesis in F. semitectum KCTC16676. Although apicidin is similar in structure to C. carbonum HC-toxin, the organization of the genes involved in apicidin biosynthesis is quite different. In C. carbonum, there are two or three copies of each of the seven HC-toxin biosynthetic genes, whereas none of the genes for apicidin biosynthesis is duplicated. With the exception of the two copies of TOXE, all copies of these seven genes in C. carbonum are dispersed throughout a ~540 kb region (Ahn et al., 2002). However, genes required for apicidin biosynthesis are clustered in a 42 kb region. Fig. 6. Northern analysis of the genes in the APS gene cluster. Total RNAs were extracted from cultures of the APS2-deleted mutant (aps2), wild-type strain KCTC16676 (WT) and APS2 overexpression mutant (APS2-OX1) grown in NPF2. The gene names are denoted along the left or right sides, and the incubation times (in days) in the medium are indicated above the blots. Ethidium bromide-stained rRNA is indicated as a loading control.

© 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466

Apicidin biosynthetic genes in Fusarium semitectum 461

Fig. 7. Proposed apicidin biosynthetic pathway in Fusarium semitectum.

Based on sequence similarity to previously characterized genes and chemical analysis of APS gene disruption mutants, we predicted the functions of individual genes and proposed an apicidin biosynthetic pathway (Fig. 7). Like other NRPSs, APS1 incorporates substrate amino acids including isoleucine (Ile), pipercolic acid (Pip) and

tryptophan (Trp). We deduced that the four APS1 substrates, in order of incorporation, are L-Pip, L-Ile (S)-Nmethoxy-Trp (Mtrp), and (S)-2-amino-8-oxodecanoic acid (Aoda) (Fig. 7). L-Pip, a non-proteineous amino acid, is an important component or precursor of many useful microbial secondary metabolites and the biosynthetic pathway

© 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466

462 J.-M. Jin et al. 䊏

for L-Pip is similar to that for lysine (He, 2006). In E. coli, D1-pyrroline-5-carboxylate (P5C) reductase, encoded by proC, catalyses the conversion of P5C into proline and also efficiently converts D1-pyrroline-6-carboxylate (P6C) into L-Pip (Fujii et al., 2002). Similarly, APS3 may be involved in the reduction of P6C to L-Pip. The presence of D-Pip instead of L-Pip in apicidin suggests that the epimerization domain of APS1 converts L-Pip to D-Pip. Based on the structure of apicidin, it is unlikely that there is an additional epimerase or racemase gene involved in apicidin biosynthesis as found in HC-toxin biosynthesis (Cheng and Walton, 2000). Mutant aps3 accumulated apicidin B instead of apicidin. Since Pip synthesis was blocked in the aps3 mutant and the APS1 enzyme has low substrate specificity, proline could be the substrate of APS1. Furthermore, low levels of apicidin B production were detected in the apicidin-overproducing mutants generated in this study and the wild-type F. pallidoroseum strain (Singh et al., 2002). APS2 was predicted to function as a transcription factor controlling apicidin biosynthesis based on the presence of a bZIP basic DNA binding motif, four ankyrin repeats, and its similarity to TOXE, a regulator of HC toxin production in C. carbonum (Ahn and Walton, 1998; Pedley and Walton, 2001). Apicidin production and the expression of the genes in the gene cluster in aps2 and APS2-OX1 mutants strongly supported this prediction. Expression patterns of the genes in the APS gene cluster are slightly different from those involved in HC-toxin biosynthesis. Although the expression of APS3, APS4, and APS11 was less influenced by APS2 than the other genes (Fig. 6), all of them were under the control of APS2. In C. carbonum, TOXA, TOXC and TOXD, but not HTS1, were regulated by TOXE (Ahn and Walton, 1998). APS5 encodes an alpha subunit of fatty acid synthase and participates in the biosynthesis of the decanoic acid backbone of Aoda. Since fatty acid synthesis in fungi requires both the alpha and beta subunits, a homologue of C. carbonum TOXC, encoding a beta subunit required for HC-toxin biosynthesis (Ahn and Walton, 1997), likely participates in apicidin production. However, a TOXC homologue is not in the APS gene cluster in F. semitectum, suggesting that a beta subunit involved in normally long chain fatty acid synthesis is utilized for apicidin production. Production of apicidin E and D2 in the aps7 and aps9 mutants, respectively, indicates that APS7 converts (S)-2-amino-decanoic acid to (2S, 8S)-2-amino-8hydroxydecanoic acid and that APS9 converts (2S, 8S)2-amino-8-hydroxydecanoic acid to (S)-2-amino-8oxodecanoic acid. We could not determine the role of APS8, which is a putative cytochrome P450 monooxygenase, but APS8 may convert decanoic acid to 2-hydroxydecanoic acid in Aoda biosynthesis. Although

the order of reactions in Aoda biosynthesis is not known, the production of apicidin E and D2 rather than apicidin in the mutants suggests that Aoda biosynthesis occurs before NRPS-catalysed reactions and APS1 has low substrate specificity. APS4 is similar to TOXF, which encodes an aminotransferase involved in HC-toxin biosynthesis. The aps4 mutant normally produced apicidin, suggesting the presence of another gene encoding an aminotransferase that converts 2-oxodecanoic acid to (S)-2-aminodecanoic acid outside of the gene cluster. Such metabolic functional redundancy is not uncommon. Disruption of a single copy of TOXF in C. carbonum produced HC-toxin, but a strain with both copies of TOXF disrupted lost HC-toxin production (Cheng et al., 1999). In Aspergillus parasiticus, two genes, norA and nor-1, have been demonstrated to encode norsolorinic acid reductases (Chang et al., 1992). However, the norA null mutant converted norsolorinic acid to averantin (Cary et al., 1999); and the nor-1 null mutant accumulated norsolorinic acid but still produced aflatoxin (Trail et al., 1994). These results suggested the presence of another enzyme that carries out the conversion of norsolorinic acid to averantin. APS6 is predicted to convert L-tryptophan to (S)-9-Nmethoxy-tryptophan because APS6 encodes an O-methyltransferase, and apicidin contains an (S)-9-Nmethoxy-tryptophan residue. APS11 is predicted to encode an efflux pump. Our inability to delete APS11 suggests that the ability to move apicidin out of the fungus or across an internal membrane is essential for survival. In a similar manner, Pitkin et al. (1996) were unable to delete both copies of the putative HC-toxin efflux pump TOXA in C. carbonum. The ability to produce apicidin is not widely distributed among Fusarium spp., including the F. semitectum–F. equiseti complex (Park et al., 1999; Jin et al., 2008). None of the fungal strains tested in this study produced apicidin, but some strains of F. asiaticum carry bands hybridized to APS genes. Analysis of differences in APS gene sequences among different species of Fusarium may provide insight in the evolution of the APS gene cluster, and the presence of APS loci in F. asiaticum suggests that some strains of this species may produce apicidin or apicidin-like metabolites. In addition, it will be our future objective to understand the transcriptional regulation mechanism of these genes, which may shed light upon the biological functions of apicidin. In conclusion, a gene cluster for apicidin biosynthesis in F. semitectum was identified, and its biosynthetic pathway was reconstructed via a combination of mutagenesis of individual genes in the pathway and chemical analysis of the resulting mutants. In addition, we were able to isolate and identify three apicidin analogues, one of which (apicidin E) is a new metabolite, and overproduce apicidin by © 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466

Apicidin biosynthetic genes in Fusarium semitectum 463

enhancing the expression of the genes in the cluster. This work establishes a foundation for gaining a deeper understanding of the mechanisms of apicidin biosynthesis and provides new opportunities to improve the production of apicidin and to produce new analogues.

Experimental procedures Strains, media and culture conditions The saprophytic fungus F. semitectum KCTC16676 isolated from soybeans in Korea (Park et al., 1999; Jin et al., 2008) was used as the wild-type strain. Fungal stock cultures were stored in 20% glycerol at -80°C and activated on potato dextrose agar (PDA). For genomic DNA isolation, fungal strains were grown in 50 ml of complete medium (CM; Leslie and Summerell, 2006) at 25°C for 72 h in a rotary shaker (150 r.p.m.). For fungal sporulation, mycelia plugs of the wildtype strain were inoculated into 50 ml of CMC (Capellini and Peterson, 1965).

Advantec no. 2 filter paper (Advantec, Dublin, CA, USA) and incubated in 70 ml of 1 M NH4Cl containing driselase (10 mg ml-1; InterSpex Products, San Mateo, CA, USA) to generate protoplasts. Transformation was performed as previously described (Kim et al., 2006).

Screening of transformants for apicidin production Vials containing 1.5 g of rice and 0.9 ml of distilled water were autoclaved at 121°C for 1 h. Individual transformants cultured on PDA were transferred into the vials and incubated at 25°C for 3 weeks. After air-drying the cultures for 5 days in a ventilated hood, 8 ml of chloroform was added to each vial. The extract was filtered through Whatman no. 2 filter papers, and 10 ml of the extract was chromatographed on plates coated with silica gel 60 (Merck, Darmstadt, Germany). Apicidin and its analogues were detected by spraying the plates with 0.5% p-anisaldehyde-sulfuric acid followed by heating at 110°C (Jin et al., 2008). Apicidin, apicidin D2 and apicidin B were purified as previously described (Jin et al., 2008) and were used as standards in the TLC assay.

Nucleic acid manipulations and PCR Fungal genomic DNA was extracted as previously described (Leslie and Summerell, 2006). For total RNA isolation, cultures of KCTC16676, aps2 and APS2-OX1 were grown in 25 ml of NPF2 (Singh et al., 2002). Total RNA was extracted from mycelia (0.1-0.2 g) using TRI reagent (Molecular Research Center, Cincinnati, OH, USA). Reverse transcriptase PCR (RT-PCR) was performed by using AccuPowerRT/PCR PreMix (Bioneer, Daejon, Korea). Standard procedures were used for plasmid DNA preparation, restriction endonuclease digestion, ligation, gel blotting and labelling of probes with 32P (Sambrook et al., 2001). PCR reactions were performed as described previously (Kim et al., 2006).

Identification and isolation of the genes involved in apicidin production A TAIL-PCR protocol was used for cloning genomic DNA flanking inserted T-DNA from apicidin deficient mutants. Genomic sequences flanking the right border of T-DNA were amplified with three arbitrary degenerate primers AD1, AD2 and AD3, and three specific primers RBn1, RBn2 and RBn3, which were positioned 224, 204 and 102 bp inside from the right end of T-DNA respectively. The tertiary TAIL-PCR product from each mutant was purified using a Geneclean Turbo kit (Qbiogene, Carlsbad, CA, USA) and cloned into pGEM-T easy vector (Promega, Madison, WI, USA) for sequencing.

ATMT and protoplast-mediated transformation

Construction of cosmid library

Agrobacterium tumefaciens-mediated transformation was performed as previously described (Mullins et al., 2001; Rho et al., 2001). KCTC16676 conidia were suspended with IM to 105-106 ml-1. A. tumefaciens strain AGl-1 that carries binary vector pKHt was grown at 28°C for 48 h in 5 ml of minimal medium and bacterial cells were diluted to an OD600 of 0.15 using IM amended with 200 mM AS. The cells were grown for an additional 6 h before mixing them with an equal volume of a spore suspension of KCTC16676. This mixture (50 ml per plate) was spread onto a nitrocellulose filter (0.45 mm pore and 45 mm diameter, Whatman) placed on co-cultivation medium. After incubation for 2 days at 25°C, nitrocellulose filters were transferred to selection medium, and hygromycin B-resistant transformants were transferred into 24-well PDA plates amended with 75 mg ml-1 hygromycin B. For protoplast-mediated transformation, fungal conidia were inoculated into 100 ml of YPG liquid medium (3 g yeast extract, 10 g peptone and 20 g glucose for 1 l) and were cultured at 25°C for 12 h in an orbital shaker (150 r.p.m.). Young mycelia were harvested by filtration through sterile

High-molecular weight chromosomal DNA of strain KCTC16676 was prepared from CM-grown mycelia according to established protocols (Bouchez and Camilleri, 1998). Purified genomic DNA (10 mg) was partially digested with MboI and treated as previously described (Orbach, 1994). The SuperCos1 vector and partially digested genomic DNA were ligated and subsequently packaged into lambda phage using a Gigapack III XL packaging extract kit (Stratagene, La Jolla, CA, USA). The packaging reaction was used to infect E. coli Xl-1 Blue MRF’Kan (Stratagene). Amplified cosmid libraries were pooled and stored at -80°C. Cosmid clones were blotted onto Hybond-N+ membranes (Amersham Biosciences, Piscataway, NJ, USA) and probed at 65°C (Sambrook et al., 2001). DNA sequencing of three cosmid clones (APS123, APS15 and FAS4) was performed by an automated sequencer (ABI Prism 3730; Applied Biosystems, Foster City, CA, USA) at GreenGene Bio Tech (Yongin, Korea). Primer walking was used to fill the gaps. Sequence analysis was performed using Lasergene v6.0 (DNASTAR, Madison, WI, USA), and the deduced amino acid sequences were blasted

© 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466

464 J.-M. Jin et al. 䊏

against the NCBI database (http://www.ncbi.nlm.nih.gov/) with the blastp algorithm.

Targeted mutagenesis, complementation and overexpression of fungal genes The candidate apicidin biosynthesis genes were deleted using the double-joint PCR (DJ-PCR) method (Yu et al., 2004). The 5′ and 3′ regions of each target gene were amplified from genomic DNA of strain KCTC16676 with primer pairs specific to these regions (Table S3). The reverse primer for the 5′ region and the forward primer for the 3′ region carried 20 bp sequence tails that overlap with both of the 5′ and 3′ ends of the 1.9 kb hygB cassette respectively. Three amplicons (5′ flanking region, hygB, and 3′ flanking region) were mixed (1:2:1 molar ratio) and used as a template for the second round of PCR, in which 10 cycles were carried out without the addition of primers, followed by an additional PCR reaction (35 cycles) with the new nested primer pairs. Resulting PCR products (approximately 5 kb long) were directly used for fungal transformation. Each transformant was transferred to a fresh PDA plate amended with 75 mg ml-1 hygromycin B and purified by a single macroconidium isolation. This purification resulted in homokaryotic transformants with homologous integration events. APS11 could not be disrupted: transformation did not yield heterokaryotic transformants with homologous integration events. For complementation of APS2, the entire APS2 gene including the native promoter and terminator was amplified with the nested PCR primer pair used in DJ-PCR (Table S3). The PCR product was directly introduced into fungal protoplasts along with vector pSK660 that carries the geneticin resistance gene (gen) as a selectable marker. For overexpression of APS2, the 5′ flanking (0.9 kb) of APS2 was amplified from wild-type strain KCTC16676 with the primers OVT-5for and OVT-5rev and was fused with hygB (1.9 kb). A 2.1 kb PCR fragment, including a putative promoter of the G. zeae b-tubulin gene, was amplified from G. zeae strain GZ03643 (Bowden and Leslie, 1999) with the primers bpro-5′ and bpro-3′ and was fused with APS2 ORF (1.3 kb) amplified from the wild-type strain with the primers OVT-5nest and OVT-3rev. These two fusion constructs were mixed for ligation and the final construct (6.2 kb) was cloned into TOPO Xl-1 vector (Invitrogen, Carlsbad, CA, USA). The plasmid containing the final construct was transformed into the wild-type strain and the expression of the APS2 gene was confirmed by Northern blot analysis.

Analysis of apicidin and its analogues Each of the mutants was grown in 30 g of rice at 25°C for 3 weeks. The rice culture was extracted with chloroform. Chloroform extract was applied to a Florisil column (2 cm inside diameter by 20 cm), washed with ethyl acetate-hexane (2:1, v/v), and then eluted with chloroform-methanol (10:1, v/v). After drying, the solute was redissolved in 10 ml of chloroform. A HP 1100 HPLC system equipped with a JEOL MS (Tokyo, Japan) was used to analyse apicidin and its analogues. The column was a Symmetry C18 column (4.6 mm ¥ 150 mm; Waters, Milford, MA, USA); the UV

detector was set at 291 nm; the mobile phase was either 75% or 85% aqueous methanol; and the flow rate was 0.2 ml min-1. To determine the chemical structures of the apicidin analogues, rice cultures (1 kg) of three mutants (aps3, aps7, and aps9) were grown at 25°C for 3 weeks. Chloroform extract was concentrated to dryness and applied to a C18 reversedphase silica column (3 cm ¥ 150 cm; Waters) and eluted with methanol-water (2:8, v/v) followed by methanol-water (7:3, v/v). The fractions containing each apicidin analogue were pooled and recrystallized from methanol to isolate 2 (apicidin E, 9 mg), 3 (apicidin D2, 12 mg) and 4 (apicidin B, 8 mg).

Survey of Fusarium species for apicidin production and presence of APS genes Thirty-eight Fusarium isolates (Table S2) (Ward et al., 2002; O’Donnell et al., 2004) were screened for apicidin production and the presence of APS genes. Genomic DNA of each strain was digested with EcoRI and fractionated in 0.7% agarose gels. Fractionated DNAs were transferred to Hybond N+ membranes (Amersham Biosciences). DNA probes were obtained by amplification of APS7, APS11 and APS1 genes using forward and reverse primers (Table S3). Genomic DNA blots were hybridized with each probe at 65°C.

Acknowledgements This work was supported by the Crop Functional Genomics Center of the 21st Century Frontier Research Program funded by the Korean Ministry of Education, Science and Technology (CG1411), and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (2009-0063350). We thank GreenGene Bio Tech for DNA sequencing and sequence analysis.

References Ahn, J.-H., and Walton, J.D. (1997) A fatty acid synthase gene in Cochliobolus carbonum required for production of HC-toxin, cyclo(d-prolyl-L-alanyl-D-alanyl-L-2-amino-9, 10-epoxi-8-oxodecanoyl). Mol Plant Microbe Interact 10: 207–214. Ahn, J.-H., and Walton, J.D. (1998) Regulation of cyclic peptide biosynthesis and pathogenicity in Cochliobolus carbonum by TOXEp, a novel protein with a bZIP basic DNA-binding motif and four ankyrin repeats. Mol Gen Genet 260: 462–469. Ahn, J.H., Cheng, Y.Q., and Walton, J.D. (2002) An extended physical map of the TOX2 locus of Cochliobolus carbonum required for biosynthesis of HC-toxin. Fungal Genet Biol 35: 31–38. Bouchez, D., and Camilleri, C. (1998) High molecular weight DNA extraction from Arabidopsis. Methods Mol Biol 82: 61–70. Bowden, R.L., and Leslie, J.F. (1999) Sexual recombination in Gibberella zeae. Phytopathology 89: 182–188. Capellini, R.A., and Peterson, J.L. (1965) Macroconidium formation in submerged cultures by a nonsporulating strain of Gibberella zeae. Mycologia 57: 962–966. © 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466

Apicidin biosynthetic genes in Fusarium semitectum 465

Cary, J.W., Barnaby, N., Ehrlich, K.C., and Bhatnagar, D. (1999) Isolation and characterization of experimentally induced, aflatoxin biosynthetic pathway deletion mutants of Aspergillus parasiticus. Appl Microbiol Biotechnol 51: 808– 812. Chang, P.-K., Skory, C.D., and Linz, J.E. (1992) Cloning of a gene associated with aflatoxin B1 biosynthesis in Aspergillus parasiticus. Curr Genet 21: 231–233. Cheng, Y.-Q., and Walton, J.D. (2000) A eukaryotic alanine racemase gene involved in cyclic peptide biosynthesis. J Biol Chem 275: 4906–4911. Cheng, Y.-Q., Ahn, J.-H., and Walton, J.D. (1999) A putative branched-chain-amino-acid transaminase gene required for HC-toxin biosynthesis and pathogenicity in Cochliobolus carbonum. Microbiology 145: 3539–3546. Darkin-Rattray, S.J., Gurnett, A.M., Myers, R.W., Dulski, P.M., Crumley, T.M., Alloco, J.J., et al. (1996) Apicidin: a novel antiprotozoal agent that inhibits parasite histone deacetylase. Proc Natl Acad Sci USA 93: 131431–113147. Desjardins, A.E. (2006) Fusarium Mycotoxins: Chemistry, Genetics and Biology. St Paul, MN: APS Press. Doekel, S., and Marahiel, M.A. (2001) Biosynthesis of natural products on modular peptide synthetases. Metab Eng 3: 64–77. Finking, R., and Marahiel, M.A. (2004) Biosynthesis of nonribosomal peptides. Annu Rev Microbiol 58: 453–488. Fujii, T., Mukaihara, M., Agematu, H., and Tsunkekawa, H. (2002) Biotransformation of l-lysine to l-pipecolic acid catalyzed by l-lysine 6-aminotransferase and pyrroline-5carboxylate reductase. Biosci Biotechnol Biochem 66: 622–627. Haese, A., Schuber, M., Herrmann, M., and Zocher, R. (1993) Molecular characterization of the enniatin synthetase gene encoding a multifunctional enzyme catalyzing N-methyldepsipeptide formation in Fusarium scirpi. Mol Microbiol 7: 905–914. Han, J.-W., Ahn, S.H., Park, S.H., Wang, S.-Y., Bae, G.-U., Seo, D.-W., et al. (2000) Apicidin, a histone deacetylase inhibitor, inhibits proliferation of tumor cells via induction of p21WAF1/Cip1 and gelsolin. Cancer Res 60: 6068–6074. He, M. (2006) Pipecolic acid in microbes: biosynthetic routes and enzymes. J Ind Microbiol Biotechnol 33: 401–407. Hong, J., Ishihara, K., Yamaki, K., Hiraizumi, K., Ohno, T., Ahn, J.W., et al. (2003) Apicidin, a histone deacetylase inhibitor, induces differentiation of HL-60 cells. Cancer Lett 189: 197–206. Jin, J.-M., Baek, S.-R., Lee, K.-R., Lee, J., Yun, S.-H., Kang, S., and Lee, Y.-W. (2008) Purification and phytotoxicity of apicidins produced by the Fusarium semitectum KCTC16676. Plant Pathol J 24: 417–422. Kim, J.-E., Jin, J., Kim, H., Kim, J.-C., Yun, S.-H., and Lee, Y.-W. (2006) GIP2, a putative transcription factor that regulates the aurofusarin biosynthetic gene cluster in Gibberella zeae. Appl Environ Microbiol 71: 1645–1652. Leslie, J.F., and Summerell, B.A. (2006) The Fusarium Laboratory Manual. Ames, IA: Blackwell Professional. Mullins, E.D., Chen, X., Romaine, P., Raina, R., Geiser, D.M., and Kang, S. (2001) Agrobacterium-mediated transformation of Fusarium oxysporum: an efficient tool for insertional mutagenesis and gene transfer. Phytopathology 91: 173– 180.

O’Donnell, K.T., Ward, T.J., Geiser, D.M., Kisler, H.C., and Aoki, T. (2004) Genealogical concordance between the mating type locus and seven other nuclear genes supports formal recognition of nine phylogenetically distinct species within the Fusarium graminearum clade. Fungal Genet Biol 41: 600–623. Orbach, M.J. (1994) A cosmid with a HyR marker for fungal library construction and screening. Gene 94: 159–162. Park, J.-S., Lee, K.-R., Kim, J.-C., Lim, S.-H., Seo, J.-A., and Lee, Y.-W. (1999) A hemorrhagic factor (Apicidin) produced by toxic Fusarium isolates from soybean seeds. Appl Environ Microbiol 65: 126–130. Pedley, K.F., and Walton, J.D. (2001) Regulation of cyclic peptide biosynthesis in a plant pathogenic fungus by a novel transcription factor. Proc Natl Acad Sci USA 98: 14174–14179. Pitkin, J.W., Panaccione, D.G., and Walton, J.D. (1996) A putative cyclic peptide efflux pump encoded by the TOXA gene of the plant-pathogenic fungus Cochliobolus carbonum. Microbiology 142: 1557–1565. Rho, H.-S., Kang, S., and Lee, Y.-H. (2001) Agrobacterium tumefaciens-mediated transformation of the plant pathogenic fungus, Magnaporthe grisea. Mol Cells 12: 407–411. Sambrook, J., Fritsch, E.F., and Maniatis, T. (2001) Molecular Cloning: A Laboratory Manual, 3nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. Scott-Craig, J.S., Panaccione, D.G., Pocard, J.A., and Walton, J.D. (1992) The cyclic peptide synthetase catalyzing HC-toxin production in the filamentous fungus Cochliobolus carbonum is encoded by a 15.7-kilobase open reading frame. J Biol Chem 267: 26044–26049. Singh, S.B., Zink, D.L., Polishook, J.D., Dombrowski, A.W., Darkin-Rattray, S.J., Schmatz, D.M., and Goetz, M.A. (1996) Apicidins: Novel cyclic tetrapeptides as coccidiostats and antimalarial agents from Fusarium pallidoroseum. Tetrahedron Lett 37: 8077–8080. Singh, S.B., Zink, D.L., Liesch, J.M., Dombrowski, A.W., Darkin-Rattray, S.J., Schmatz, D.M., and Goetz, M.A. (2001) Structure, histone deacetylase, and antiprotozoal activities of apicidins B and C, congeners of apicidin with proline and valine substitutions. Org Lett 3: 2815– 2818. Singh, S.B., Zink, D.L., Liesch, J.M., Mosley, R.T., Dombrowski, A.W., Bills, G.F., et al. (2002) Structure and chemistry of apicidins, a class of novel cyclic tetrapeptides without a terminal alpha-keto epoxide as inhibitors of histone deacetylase with potent antiprotozoal activities. J Org Chem 67: 815–825. Trail, F., Chang, P.-K., Cary, J., and Linz, J.E. (1994) Structural and functional analysis of the nor-1 gene involved in the biosynthesis of aflatoxins by Aspergillus parasiticus. Appl Environ Microbiol 60: 4078–4085. Ueda, T., Takai, N., Nishida, M., Nasu, K., and Narahara, H. (2007) Apicidin, a novel histone deacetylase inhibitor, has profound anti-growth activity in human endometrial and ovarian cancer cells. Int J Mol Med 19: 301–308. Ward, T.J., Bielawski, J.P., Kistler, H.C., Sullivam, E., and O’Donnell, K. (2002) Ancestral polymorphism and adaptive evolution in the trichothecene mycotoxin gene cluster of phytopathogenic Fusarium. Proc Natl Acad Sci USA 99: 9278–9283.

© 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466

466 J.-M. Jin et al. 䊏

Weber, G., Schörgendorfer, K., Schneider-Scherzer, E., and Leitner, E. (1994) The peptide synthetase catalyzing cyclosporine production in Tolypocladium niveum is encoded by a giant 45.8-kilobase open reading frame. Curr Genet 26: 120–125. Yu, J.-H., Hamari, Z., Han, K.-H., Seo, J.-A., ReyesDominguez, Y., and Scazzocchio, C. (2004) Double-joint PCR: a PCR-based molecular tool for gene manipulations in filamentous fungi. Fungal Genet Biol 41: 973–981.

Supporting information Additional supporting information may be found in the online version of this article. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

© 2010 Blackwell Publishing Ltd, Molecular Microbiology, 76, 456–466