Molecular Basis for Chloronium-mediated Meroterpene Cyclization

Structural inspection of the bacterial meroterpenoid antibiotics belonging to the napyradiomycin family of chlorinated dihydroquinones suggests that the biosynthetic cyclization of their terpenoid subunits is initiated via a chloronium ion. The vanadium-dependent haloperoxidases that catalyze such reactions are distributed in fungi and marine algae and have yet to be characterized from bacteria. The cloning and sequence analysis of the 43-kb napyradiomycin biosynthetic cluster (nap) from Streptomyces aculeolatus NRRL 18422 and from the undescribed marine sediment-derived Streptomyces sp. CNQ-525 revealed 33 open reading frames, three of which putatively encode vanadium-dependent chloroperoxidases. Heterologous expression of the CNQ-525-based nap biosynthetic cluster in Streptomyces albus produced at least seven napyradiomycins, including the new analog 2-deschloro-2-hydroxy-A80915C. These data not only revealed the molecular basis behind the biosynthesis of these novel meroterpenoid natural products but also resulted in the first in vivo verification of vanadium-dependent haloperoxidases.

Nature has devised several mechanisms to polarize the terminal olefin of linear terpenes to facilitate the creation of new C-X bonds. For instance, cyclization of the C 30 hydrocarbon squalene to steroids and hopanoids is initiated, respectively, by epoxidation or protonation of the terminal olefin. Although these biosynthetic strategies are widely distributed, a third mechanism for terpene cyclization has been characterized in marine macroalgae involving bromonium ion-induced ring closure (1)(2)(3). Oxidation of the halide is catalyzed by vanadiumdependent bromoperoxidase in the presence of hydrogen peroxide to produce the corresponding hypohalous acid. This species then further reacts with electron-rich organic substrates in a regio-and stereoselective manner, giving rise to brominated terpenes and other halogenated natural products (1,2,4). Vanadium-dependent bromoperoxidases are widely distributed in marine algae, and the first enzyme was discovered in 1984 from the brown alga Ascophyllum nodosum (5).
Vanadium chloroperoxidases (V-ClPOs), 3 on the other hand, have been isolated primarily from dematiaceous hyphomycete fungi (1). The first enzyme was characterized in 1993 from Curvularia inaequalis (6), and even though there are numerous chlorinated marine natural products, V-ClPOs have not been reported from marine organisms to date (1,2,7). Although the biological function of V-ClPOs has not yet been elucidated, marine algal vanadium-dependent bromoperoxidases have been shown through in vitro chemoenzymatic conversions to catalyze bromonium ion-initiated cyclization of terpenes and ethers (1,8). These studies not only demonstrated that the enzymes were able to initiate cyclization of a terpene by a bromonium ion but also proved that the halogenation reaction occurred with stereochemical control. To date, all known V-dependent haloperoxidases have been characterized in vitro from eukaryotic systems (9).
Recent biosynthetic gene cluster analysis of the hybrid monoterpene-polyketides furaquinocin from Streptomyces sp. strain KO-3988 and furanonaphthoquinone I from Streptomyces cinnamonensis DSM 1042 revealed a clustering of genes coding for the type III polyketide synthase THN synthase, a prenyltransferase, and a complete set of mevalonic acid-specific geranyl diphosphate biosynthetic enzymes (19,20). Herein we report the cloning and sequence analysis of the napyradiomycin biosynthetic gene cluster (nap) from Streptomcyces sp. CNQ-525 and S. aculeolatus NRRL 18442, its functional expression in the heterologous host Streptomyces albus, and the structure elucidation of a new napyradiomycin analog. The bioinformatics analysis of the nap cluster, which contains the unprecedented distribution of three V-ClPOs, further revealed the molecular basis for the chlorination and cyclization of the terpene units in the napyradiomycin family of antibiotics.

MATERIALS AND METHODS
Bacterial Strains, Plasmids, and Culture Conditions-Streptomyces sp. CNQ-525 (14) was a gift from Dr. William Fenical (Scripps Institution of Oceanography). For sporulation, it was grown in A1 growth medium for nine days at 30°C with shaking (250 revolutions/min). pOJ446 (21) was used for cosmid library construction and was grown in LB liquid medium. The expression host S. albus was provided by Dr. Joern Piel (University of Bonn). Strain and DNA manipulations as well as southern hybridization experiments were performed according to standard procedures (22,23).
Cosmid Library Construction and Sequencing-pOJ446 was digested with HpaI, dephosphorylated with calf intestinal phosphatase, and digested with BamHI. Streptomyces sp. CNQ-525 genomic DNA was partially digested with Sau3AI. DNA fragments between 35 and 40 kb were isolated by agarose gel electrophoresis and purified for ligation. Before adding the ligase, the mixture containing plasmid and genomic DNA was incubated at 50°C for 5 min and then placed on ice for 5 min. Ligation was carried out overnight at 16°C using T4 DNA ligase. Packaging and titration were performed using XL-1 blue MRF cells (Stratagene) and MaxPlax packaging extracts (Epicenter) according to the manufacturer's instructions. After selection with apramycin (100 g/ml), ϳ2000 colonies were picked and grown overnight in 96-well plates. Cultures were transferred to a Hybond-N membrane (Amersham Biosciences) and grown overnight on LB agar containing apramycin (100 g/ml). Colonies were lysed and DNA denatured and fixed according to standard procedures (22). The library was screened using Streptomyces sp. CNH-099-based type III polyketide synthase and prenyltransferase genes as heterologous probes. 4 Peroxidase labeling and detection of enzyme-la-  beled positives were performed with the ECL direct nucleic acid labeling and detection systems kit (Amersham Biosciences). Cosmid clone pJW6F11 was sequenced by the shotgun method (Macrogen Inc., Seoul, Korea) and annotated with BLASTP (24) and FRAMEPLOT (25). The genome scanning of S. aculeolatus NRRL 18422 was previously reported (26). The GenBank TM accession numbers for the nap cluster are EF397639 from Streptomyces sp. CNQ-525 and EF397638 from S. aculeolatus.  Streptomyces Transformation, Culture Conditions, and MS Analysis of Halogenated Meroterpenoids-pJW6F11 was introduced into S. albus protoplasts, apramycin-resistant transformants were selected, and the transformants were cultivated according to standard procedures (23). Several 100-ml cultures of S. albus/pJW6F11 were grown in R2YE containing 100 g/ml apramycin at 30°C for 9 -14 days by rotary shaking (250 revolutions/min). The cells were lysed using 80% MeOH:H 2 O, and after extraction with 1:1 MeOH:CH 2 Cl 2 , the production of nap-based compounds was analyzed using a Hewlett Packard 1100 series high performance liquid chromatography system linked to an Agilent ESI-1100 MSD mass spectrometer (gas flow set to 13 ml/min, drying temperature set to 350°C, and nebulizing pressure set to 40 pounds/square inch). A Luna 4.6 ϫ 150 mm C18 column was used at a flow rate of 0.7 ml/min with a linear solvent gradient of 10 -100% acetonitrile in water over a period of 20 min. Fourier transform mass spectral analysis was accomplished with a LTQ-FTMS (ThermoFinnigan). Diluted organic extracts from S. albus/pJW6F11 were introduced into the spectrometer via direct infusion at 1-3 l/min. The sheath gas was set to 4 liters/min, and the capillary inlet was set to 275°C. The signal was optimized at the 509 m/z peak using the autotuning feature in the LTQ portion of the instrument. All FTMS analysis was performed in the negative ion mode, and the data were collected at 200,000 resolution. Extracts from clones harboring pOJ446 without insert did not produce any napyradiomycins. All of the theoretical values were obtained by importing the molecular formulas into the Qual browser software (ThermoFinnigan).
Purification of 2-Deschloro-2-hydroxy-A80915C-Cultures of the wild-type Streptomyces sp. CNQ-525 (14) were grown in A1 growth medium for nine days at 30°C with shaking (250 revolutions/min) and then extracted with EtOAc. Extracts were dried over anhydrous MgSO 4 and concentrated in vacuo. The crude extract (310 mg) from a 1-liter fermentation was subjected to reversed phase C 18 18 analytical LC-MS as described above, and the fraction eluting with 3:2 MeCN/H 2 O was subjected to purification by high performance liquid chromatography using a Waters differential refractometer R401 detector. Compounds were purified on a Luna 250 ϫ 10 mm C8 column employing an isocratic condition of 67% MeCN/H 2 O with a flow rate of 2.0 ml/min. 2-Deschloro-2-hydroxyl-A80915C (4, 4.3 mg) eluted between 17 and 20 min. NMR spectra were recorded on a Varian Inova 500-MHz spectrometer. 1 H and 13 C chemical shifts were referenced to the solvent peak (CDCl 3 ) ␦ 7.26 and 77.0, respectively. Standard parameters were used for one-and two-dimensional NMR spectra. 1

RESULTS AND DISCUSSION
The napyradiomycin biosynthetic cluster (nap) was discovered from each bacterium using distinct methods. In Streptomyces sp. CNQ-525, PCR-amplified THN synthase and prenyltransferase gene fragments were used as probes for the identification of a single pOJ446 cosmid clone (pJW6F11) containing a 36-kb genomic insert, which was sequenced by a shotgun approach. Identification of the complete 43-kb nap locus was alternatively achieved in S. aculeolatus NRRL 18422 by genome scanning (26). When aligned, the two clusters are similarly organized and 97% iden- tical at the nucleotide level (Fig. 2). The DNA sequence is interrupted by a ϳ350-bp gap between the convergent genes napH2 and napT8 that proved impervious to our sequencing efforts in CNQ-525. Analysis of the 43-kb nap cluster revealed 33 open reading frames, which included five genes putatively involved in the construction of the naphthoquinone polyketide core (napB1-B5), nine genes associated with the biosynthesis (napT1-T7) and attachment (napT8 -T9) of the terpenoid units, four halogenases (napH1-H4), nine putative regulatory and resistance proteins (napR1-R9), four open reading frames of unknown function (napU1-U4), and two transposases that suggest this cluster may have been acquired via horizontal gene transfer (Table 1 and Fig.  2). Of the four nap halogenases, three show striking similarity to fungal V-ClPOs and a hypothetical protein TioM from Micromonospora sp. ML1, which is unprecedented in prokaryotic gene clusters.
To verify that all of the genetic information for the production of chlorinated dihydroquinones, such as A80915C (1), 3Љ-hydroxy-7-methylnapyradiomycin B2 (2), and 2-deschloro-A80915C (3), is contained on the nap gene cluster, the Streptomyces sp. CNQ-525 cosmid clone pJW6F11 that contained the truncated 36-kb nap cluster was heterologously expressed in S. albus. Although the mevalonic acid pathway genes napT1-T4 and part of napT5 are absent in pJW6F11, the host strain was reasoned to provide the isopentenyl pyrophosphate and dimethylallyl pyrophosphate building blocks from primary metabolism. Co-injection with authentic napyradiomycin standards, high performance liquid chromatography-mass spectrometry, and FTMS analysis of the organic extracts from S. albus/pJW6F11 unequivocally demonstrated that the transformant yielded the nap-based chlorinated dihydroquinones 1-3 as well as a series of related analogs that exhibited the characteristic napyradiomycin chromophore (14) with high UV light absorption at 258, 300, and 356 nm (Fig. 3). The observed masses for two of these dichlorinated analogs at m/z 527 and 509 (negative ion mode) were observed in both the wild type strain CNQ-525 and the S. albus/pJW6F11 transformant, whereas a monochlorinated analog (m/z 473) and a dichlorinated analog (m/z 525) were only identified in the transformant.
High  (Fig. 4). The unknown dichlorinated 509 species (c) observed by high performance liquid chromatography-mass spectrometry analysis (Fig. 3) likely has the same molecular formula as compound 2 (C 26 H 31 Cl 2 O 6 Ϫ ), so its exact mass could not be distinguished by FTMS.
Fermentation of Streptomyces sp. CNQ-525 followed by extraction and chromatography provided the dichlorinated 527 species (4) in 4.3 mg/liter. Analysis of the proton and carbon NMR spectra, with the aid of gradient-enhanced heteronuclear multiple bond correlation data, clearly established that 4 contained the chlorocyclohexyl monoterpenoid unit common to 1-3. NMR comparison with structurally related compounds 1 and 3 indicated that they only differed in the substitution at C-2. The presence of a broad singlet at ␦ 4.16 in the 1 H spectrum for 4 indicated an additional hydroxyl signal not observed in 1 and 3, which was confirmed by high resolution FTMS. This new C-2 hydroxyl substitution resulted in subtle differences in the NMR spectra of 4 in comparison to that of the chloro analog 1 at C-1Ј, in which the methylene protons shift from ␦ 2.62 and 2.45 in 1 to ␦ 2.15 in 4. The relative stereochemistry of 4 was assigned by comparing nuclear Overhauser effect spectroscopy correlations to those previously described for 1 (14,27). Expression of the nap cluster in S. albus unequivocally confirmed that all of the genes required for the production of the 7-methylnapyradiomycin family of chlorinated dihydroquinones were contained on the cosmid clone pJW6F11. A biosynthetic scheme for the production of napyradiomycins 1-3 was proposed on the basis of the molecular logic of the nap cluster ( Fig. 1). As in furaquinocin (19) and furanonaphthoquinone I (20) biosynthesis, assembly of the dihydroquinone core is catalyzed by the type III polyketide synthase homologous THN synthase NapB1, which condenses five malonyl-CoA molecules to THN. The monooxygenase NapB2 putatively oxidizes THN to flaviolin, as is the case in S. antibioticus IF013271 (28), which is then methylated by the methyltransferase NapB5. Attachment of the first isoprene unit, using dimethylallyl pyrophosphate as the substrate, occurs through a nucleophilic attack involving one of the two prenyltransferases (NapT8/T9). Hydrogenation or chlorination at C-2 by the FADH 2 -dependent halogenase NapH2 may facilitate the second prenylation reaction at C-3 with geranyl pyrophosphate by the second nap prenyltransferase to yield the diprenylated intermediate SF2415B1 (13). Cyclization of the hemiterpene subunit via a chloronium ion is putatively catalyzed by one of the three nap V-ClPOs (NapH1, H3, H4) giving rise to the 7-methyl derivative of napyradiomycin A1. The monoterpene subunit of 7-methylnapyradiomycin A1 putatively undergoes a related V-ClPO-facilitated cyclization to form 1 and 3, in which a molecule of water is incorporated into the cyclohexanol moiety. Further dechlorination of 1 likely gives rise to 2 (Fig. 1). Although the specifics of the biosynthetic reaction sequence have yet to be confirmed, the suc-cessful heterologous expression of the nap cluster confirms that all of the genes required for biosynthesis are present. Phylogenetic analysis of the three nap V-ClPOs with characterized fungal V-ClPOs and algal vanadium-dependent bromoperoxidases shows that the three nap V-ClPOs form their own clade, with NapH1 and NapH4 being most closely related (Fig. 5A). Inspection of the amino acid residues in the active sites of the vanadium haloperoxidases suggests that NapH3 may not be directly involved in the chlorination and cyclization of the terpene units and may instead act as a hydroxylase. Structural and functional comparisons of vanadium haloperoxidases from eukaryotic organisms have shown that six amino acids are required for halogenating activity, whereas two are required for halide selectivity (29 -32). All three nap enzymes contain His-496 (which covalently binds to the vanadate co-factor) and five of the six residues that participate in hydrogen bonding, Lys-353, Arg-360, Ser-402, Gly-403, and Arg-490. Conserved residue His-404, which is proposed to form a hydrogen bond to the apical oxygen of the co-factor (29), is replaced with Ser (as seen in NapH1 and NapH4) or Phe (as in NapH3). Mutagenesis studies have shown that the V-ClPO mutant H404A from C. inaequalis (29) loses chlorinating activity, and a similar natural exchange is observed in NapH3. Hence, NapH3 may not be involved in halogenation but rather may be used to hydrate aliphatic carbons to produce such compounds as 7-methylnapyradiomycin A2 (15). NapH1 and NapH4, on the other hand, contain a hydrophilic Ser residue at this position and are there- FIGURE 5. A, relatedness of NapH1, NapH3, NapH4, and other vanadium haloperoxidases from fungi and algae. Phylogenetic analysis was performed using ClustalW (33), and the unrooted tree was visualized by TreeView. The scale bar indicates 0.2 changes/amino acid. B, alignment of known vanadium haloperoxidases to NapH1, NapH3, and NapH4 around the conserved active site residues, which are highlighted in light gray, whereas the differences are highlighted in dark gray. Sequence identification codes include Ci_VClPO from C. inaequalis (GenBank TM accession number CAA59686); Rb_VClPO from Rhodopirellula baltica SH1 (CAD72609); Ed_VClPO from Embellisia didymospora (CAA72622); An_VBrPO from A. nodosum (P81701); Co_VBrPO from Corallina officinalis (AAM46061); and Cp_VBrPO from Corallilna pilulifera (BAA31261). fore predicted to catalyze the chloronium-induced cyclization of the two terpene units. Analysis of the recombinant proteins is presently underway and will help delineate their functions and specificities.
In conclusion, we have demonstrated for the first time the molecular basis for the chlorination and cyclization of terpene units involving novel bacterial V-ClPOs. The isolation of the nap biosynthetic cluster from two bacteria provides a powerful toolbox to study these unique halogenating enzymes.