DOI: 10.1002/cbic.201500040
Communications
Genome Mining of the Hitachimycin Biosynthetic Gene Cluster: Involvement of a Phenylalanine-2,3-aminomutase in Biosynthesis Fumitaka Kudo,[a] Koichi Kawamura,[a] Asuka Uchino,[a] Akimasa Miyanaga,[a] Mario Numakura,[a] Ryuichi Takayanagi,[b] and Tadashi Eguchi*[b] Hitachimycin is a macrolactam antibiotic with (S)-b-phenylalanine (b-Phe) at the starter position of its polyketide skeleton. To understand the incorporation mechanism of b-Phe and the modification mechanism of the unique polyketide skeleton, the biosynthetic gene cluster for hitachimycin in Streptomyces scabrisporus was identified by genome mining. The identified gene cluster contains a putative phenylalanine-2,3-aminomutase (PAM), five polyketide synthases, four b-amino-acid-carrying enzymes, and a characteristic amidohydrolase. A hitA knockout mutant showed no hitachimycin production, but antibiotic production was restored by feeding with (S)-b-Phe. We also confirmed the enzymatic activity of the HitA PAM. The results suggest that the identified gene cluster is responsible for the biosynthesis of hitachimycin. A plausible biosynthetic pathway for hitachimycin, including a unique polyketide skeletal transformation mechanism, is proposed.
the starter position of their polyketides (Scheme 1). Thus, common modification enzymes are proposed to be responsible for this unique transformation. Although the cremimycin biosynthetic gene cluster has recently been identified and several unassigned genes are speculated to be involved in the unique polyketide modification,[6] the modification mechanism remains unknown. Thus, identification of comparable biosynthetic genes should provide insight into the unique multi-step enzymatic polyketide modification mechanism. Although the mechanism of b-Phe incorporation into the polyketide during hitachimycin synthesis is unclear, it is likely to be similar to related biosynthetic pathways for macrolactam antibiotics, including those for cremimycin, incednine, and vicenistatin. To further our knowledge of b-amino acid biosynthetic machinery and selective incorporation mechanisms, this study aimed to identify the hitachimycin biosynthetic gene cluster. This knowledge will help in the development of rational biosynthesis protocols for this class of polyketide. We used a genome-scanning approach to identify the hitachimycin biosynthetic gene cluster in the producer strain S. scabrisporus JCM 11712. Chromosomal DNA was extracted and sequenced with a combination of an Illumina HiSeq 2000 and a Roche Genome Sequencer FLX (mate pair sequencing analysis) to obtain ~ 11 Mb of draft genome sequence. In silico screening with the functionally characterized Taxus chinensis PAM (involved in taxol biosynthesis)[7] showed a contiguous gene cluster containing genes for a putative PAM, five polyketide synthases (PKSs), four characteristic b-amino acid-carrying enzymes, and one amidohydrolase (Table 1, Scheme 2). No other candidate gene cluster with these features was found in the genome sequence of JCM 11712. Thus, we named this the “hit” gene cluster (hitachimycin biosynthesis). During preparation of this manuscript, the draft genome sequence of a similar strain (S. scabrisporus DSM 41855) was deposited; this contains the same genes for hitachimycin biosynthesis as JCM 11712; however, several PKS genes, such as hitP2 and hitP3, are absent from the sequence data in the draft genome sequence of S. scabrisporus DSM 41855. Thus, our sequence analysis provides a more complete sequence of the hitachimycin biosynthetic gene cluster. Based on bioinformatic analysis of the genes, we propose a pathway for hitachimycin biosynthesis (Scheme 2). First, the putative PAM HitA converts l-Phe to (S)-b-phenylalanine because of the stereochemistry at C-21 of hitachimycin. (S)-b-Phe is recognized by a b-amino-acid-selective ATP-dependent ligase (HitB) and transferred to the discrete acyl carrier protein
Hitachimycin (also known as stubomycin) was isolated from Streptomyces scabrisporus and shown to be an antiprotozoal and antitumor antibiotic, as well as having potent antimicrobial activity against Bacillus subtilis and Micrococcus luteus.[1] However, its low solubility (in any solvent) has prevented examination of its precise mode of action. The main structural features of hitachimycin are the unique b-phenylalanine (b-Phe) starter unit of its polyketide skeleton and a bicyclic structure with an internal five-membered carbocycle (Scheme 1).[2] Incorporation studies illustrated that the bPhe starter unit is derived from a-phenylalanine, presumably by the action of phenylalanine-2,3-aminomutase (PAM).[2] To construct the five-membered carbocyclic moiety within the macrocyclic skeleton, unknown enzymes encoded in the biosynthetic gene cluster modify the elongated polyketide chain or macrocycle with the b-Phe starter unit. Similar five-membered cyclic structures are found in cremimycin,[3] piceamycin,[4] and viridenomycin,[5] which also have b-amino acids at [a] Prof. Dr. F. Kudo, K. Kawamura, A. Uchino, Dr. A. Miyanaga, M. Numakura Department of Chemistry, Tokyo Institute of Technology 2-12-1 O-okayama, Meguro-ku, Tokyo 152-8551 (Japan) [b] R. Takayanagi, Prof. Dr. T. Eguchi Department of Chemistry and Materials Science Tokyo Institute of Technology 2-12-1 O-okayama, Meguro-ku, Tokyo 152-8551 (Japan) E-mail: [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201500040.
ChemBioChem 0000, 00, 0 – 0
These are not the final page numbers! ÞÞ
1
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Communications
Scheme 1. Hitachimycin and related b-amino-acid-containing macrolactams.
(ACP), HitD, to form (S)-b-Phe-HitD. Then, another ATP-dependent ligase HitE activates a certain amino acid, which is transferred to the amino group of (S)-b-Phe-HitD to form aminoacyl-(S)-b-Phe-HitD. The aminoacyl-(S)-b-Phe moiety of dipeptidyl-HitD is transferred to a loading ACP domain of PKS HitP1 by the aminoacyltransferase HitC. This incorporation of (S)-b-Phe onto the PKS assembly line is same as in the vicenistatin biosynthetic pathway, except for the decarboxylation step.[8] Based on the domain structures and substrate specificity of the acyltransferase (AT) domain of the PKSs, HitP1, HitP2, HitP3, HitP4, and HitP5, a rational PKS assembly line was predicted (Scheme 2). Although the mechanism of formation of the C6=C7 bond with presumably Z stereochemistry cannot be predicted from the sequence, the other predictions for the polyketide chain were consistent with the chemical structure of hitachimycin (alignment of PKS domains in Figure S1 in the Supporting Information). The terminal aminoacyl group in the elongated polyketide chain is removed by the amidohydrolase HitF, followed by macrocyclization catalyzed by the thioesterase (TE) domain in the PKS, HitP3. The macrocyclic polyketide released from the PKS assembly line appeared to be modified by post-PKS modification enzymes. Interestingly, five genes (hitM1, hitM2, hitM3, hitM4, and hitM5) are conserved between hitachimycin and cremimycin biosynthetic gene clusters; this indicated that these genes are involved in forming the characteristic five-membered carbocycle of the polyketide skeleton (Figure S2).[6] Because HitM1 and HitM4 are putative short-chain dehydrogenases/reductases, and HitM3 is a putative cytochrome P450, oxidation reactions likely trigger the C C bond formation between C8 and C12 of the polyketide chain. The hydroxy group at C-11 is oxidized by either HitM1 or HitM4. It appears that HitM3 (cytochrome P450) catalyzes hydroxylation at C-10 and converts the hydroxy group to a carbonyl group. The presumptive a,b-unsaturated
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
ketone intermediate would be converted to a five-membered carbocycle by Michael addition or a Nazarov cyclization mechanism, possibly by HitM2 and HitM5 (two unassigned putative sugar-isomerase-type enzymes), which might act as acid/base catalysts. The highly oxidized intermediate is then reduced by either HitM1 or HitM4, and finally the putative methyltransferase HitM6 completes the biosynthesis of hitachimycin by methylation of the hydroxy group at C-10. The oxidation states of the biosynthetic intermediates generated by the redox enzymes are likely to be important for understanding this cyclization mechanism. The hit gene cluster also contains four putative regulatory genes (hitR1, hitR2, hitR3, and hitR4) and a putative transporter gene (hitT1). Thus, it appears that the identified hit gene cluster is responsible for the biosynthesis of hitachimycin. To confirm that this gene cluster is responsible for hitachimycin biosynthesis, we knocked out hitA by inserting an apramycin-resistance gene cassette (by using PCR-based gene replacement for Streptomyces genome manipulation).[9] The obtained DhitA mutant did not produce hitachimycin, thus indicating that this gene is involved in hitachimycin synthesis (Figure 1 A). When we supplemented the DhitA strain with (S)-bPhe, hitachimycin production was restored (Figures 1 B and S3). In contrast, supplementation with (R)-b-Phe did not lead to production of a corresponding diastereomer of hitachimycin or any related compound (Figure 1 C); presumably the stereochemistry of b-Phe is critical for recognition by the b-aminoacid-selective ATP-dependent ligase HitB or downstream biosynthetic enzymes. The stereochemistry at C-21 might be critical for macrocyclization of the polyketide chain. These results illustrate that hitA is involved in hitachimycin biosynthesis, and that the protein product is a PAM that provides (S)-b-Phe. Next, we examined the PAM enzymatic activity of HitA in vitro. Recombinant HitA was expressed in Escherichia coli and 2
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Communications Table 1. Hitachimycin biosynthesis genes and putative functions. Gene
Size [bp] (aa)
Accession numbers of encoded protein homologues Positives [%]/ Identity [%]
hitM1 hitM2 hitA hitP1
870 864 1551 11949
WP_020550023 WP_020550024 WP_026218103 WP_020550026
(290) (288) (517) (3983)
(DSM 41855),[b] (DSM 41855),[b] (DSM 41855),[b] (DSM 41855),[b]
cmiM1 (BAO66515), 71/79 cmiM2 (BAO66516),71/80 encP (AAF81735), 73/85 PKS, WP_014060784
hitB
1599 (533)
WP_026218104 (DSM 41855),[b] MlaJ (ACO94489), 61/73
hitC hitD hitE
945 (315) 246 (82) 1548 (516)
WP_020550028 (DSM 41855),[b] CmiS5, BAO66532, 60/71 WP_020550029 (DSM 41855),[b] MlaS, ACO94491, 57/70 WP_020550030 (DSM 41855),[b] MlaL, ACO94492, 64/73
hitR1
720 (240)
hitM3 hitM4 hitR2
1236 (412) 864 (288) 435 (145)
hitT1
1614 (538)
hitP2 hitF hitP3 hitP4
9975 906 6012 15558
(3325) (302) (2004) (5186)
hitP5 hitM5 hitM6
4836 (1612) 903 (301) 651 (217)
hitR3 hitR4
2652 (884) 2802 (934)
WP_020550031 (DSM 41855),[b] Streptomyces katrae, TetR family transcriptional regulator, WP_030302823, 47/63 WP_020550032 (DSM 41855),[b] cmiM4 (BAO66524), 73/84 WP_020550033 (DSM 41855),[b] cmiM3 (BAO66523), 71/81 WP_026218105 (DSM 41855),[b] cold-shock protein, (WP_018381685), 45/58 WP_026218106 (DSM 41855),[b] MFS transporter, (WP_030976952), 50/69 PKS, (WP_014060789), 56/66 WP_020549776 (DSM 41855),[b] cmiM6 (BAO66540), 75/85 cmiP6 (BAO66541), 65/75 WP_020550052 (DSM 41855),[b] cmiP7 (BAO66542), 67/76 WP_026219432 (DSM 41855),[b] cmiP8 (BAO66543), 65/75 WP_026219433,[b] cmiM7(BAO66544), 69/83 WP_020556451 (DSM 41855),[b] O-methyltransferase (EYF02720), 49/67 cmiR5 (BAO66545), 47/58 Orf5, (BAO66546), 47/60
Putative function
Homologues in vin, cmi, and idn clusters[c] (except for PKS)
short-chain dehydrogenase sugar phosphate isomerase/epimerase phenylalanine-2,3-aminomutase PKS (ACP-KS-AT-DH-KR-ACP-KS-AT-DHER-KR-ACP) ATP-dependent b-aminoacyl-ACP synthetase ACP aminoacyltransferase ACP ATP-dependent aminoacyl-ACP synthetase TetR family transcriptional regulator
cmiM1 cmiM2
cytochrome P450 monooxygenase 3-ketoacyl-ACP reductase cold-shock protein (regulator)
vinN, cmiS6, idnL1 vinK, cmiS5, idnL2 vinL, cmiS4, idnL6 vinM, cmiS3, idnL7
cmiM4 cmiM3
MFS transporter PKS (KS-AT-KR-ACP-KS-AT-DH-KR[a]-ACP) amidohydrolase PKS (KS-AT-DH-KR-ACP-TE) PKS (KS-AT-DH-KR-ACP-KS-AT-DHKR-ACP-KS-AT-DH-KR-ACP) PKS (KS-AT-KR-ACP) sugar phosphate isomerase/epimerase O-methyltransferase LuxR-family transcriptional regulator LuxR-family transcriptional regulator
vinJ, cmiM6, idnL5
cmiM7
cmiR5
[a] This KR domain appeared to be inactive because of the lack of a catalytic tyrosine and NADPH-binding motif. [b] These genes are in the draft genome sequence of S. scabrisporus DSM 41855. [c] vin: vicenistatin biosynthetic gene; cmi: cremimycin biosynthetic gene; idn: incednine biosynthetic gene. PKS: polyketide synthase, ACP: acyl carrier protein, KS: b-ketosynthase, AT: acyltransfearse, DH: dehydratase, ER: enoylreductase, KR: b-ketoreductase, TE: thioesterase.
Figure 1. Inactivation of hitA. HPLC analysis (detection at 300 nm) of the crude extracts from the indicated cultures.
indicating that HitA has a strict stereoselective PAM reaction mechanism. This enzymatic activity of HitA supports the hypothesis that the hit gene cluster is responsible for hitachimycin biosynthesis. In summary, we identified the hitachimycin biosynthesis gene cluster (hit) by genome mining, and characterized the (S)b-Phe-forming PAM, HitA. The identified hit gene cluster is
purified by affinity chromatography (see the Supporting Information). As expected, HitA specifically converted l-Phe to (S)b-Phe (Figure 2 A). The reverse reaction (from (S)-b-Phe to lPhe) was also observed (Figure 2 C). We confirmed the stereochemistry of the HitA reaction product to be the S form after 2,4-dinitrophenyl-5-l-alanine amide derivatization with Marfey’s reagent (Figure S5). Neither d-Phe nor (R)-b-Phe reacted, thus ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
These are not the final page numbers! ÞÞ
3
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Communications
Scheme 2. Hitachimycin biosynthetic gene cluster and the proposed biosynthesis pathway for hitachimycin.
Experimental Section
a novel member of the b-amino-acid-containing macrolactam biosynthesis group. The presence of both b-amino-acid-forming and b-amino-acid-carrying enzymes is critical for identification of this family of macrolactam polyketide biosynthetic gene clusters. In fact, a wide variety of natural products, such as isoprenoids and polyketides, are obtained by combinations of PAMs and other enzymes involved in metabolism.[10] The accumulation of knowledge of biosynthetic genes and enzymes is important to allow manipulation of the biosynthetic machinery to obtain new compounds. Mutasynthesis with various structurally similar b-amino acids by using the DhitA strain is now in progress. However, to achieve rationally engineered biosynthesis, the substrate-recognition mechanisms of HitB, HitC, HitE, and PKS enzymes should be elucidated before they can be altered.
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
Genome sequence analysis: S. scabrisporus JCM 11712 chromosomal DNA was extracted by standard phenol/chloroform extraction.[11] DNA sequencing and assembly of the obtained DNA sequence data were conducted by Takara. Automated gene annotation was performed with the MiGAP server (http://www.migap.org/ index.php/en), and the annotated open reading frames were revised by FramePlot analysis (http://www0.nih.go.jp/ ~ jun/cgi-bin/ frameplot.pl). In silico screening with the T. chinensis PAM gene and routine genetic analyses were conducted by using Geneious (Biomatters, Auckland, New Zealand). The 80-kb DNA sequence including the putative hitachimycin biosynthetic gene cluster was deposited in the DDBJ under the accession number LC008143. Disruption of hitA: A pOJ446-based cosmid library in Escherichia coli XL1 Blue MRF’ was constructed from chromosomal DNA. Several hit-gene-containing cosmids were selected by colony hybridiza-
4
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Communications cassette) were used to introduce an apramycin gene cassette into the middle of hitA. A DNA fragment (1384 bp) from plasmid pIJ773 (Plant Bioscience Limited, Norwich, UK) digested with HindIII and EcoRI was used as a template to prepare the hitA disruption DNA fragment. The PCR reaction mixture (10 mL) comprised PrimeSTAR GXL reaction buffer (2 mL), dNTPs (0.8 mL, 2.5 mm), primers (0.2 mL, 10 mm each), template (3 mL, 13 ng mL 1), and PrimeSTAR GXL polymerase (0.2 mL; Takara). Thermal cycling conditions were: 10 cycles of 94 8C for 45 s, 50 8C for 45 s, and 72 8C for 90 s, followed by 15 cycles of 94 8C for 45 s, 55 8C for 45 s, and 72 8C for 90 s. The amplified DNA fragment was introduced into BW25141/pKD78/ hitA-pUC119 by electroporation and the desired transformant was cultured in LB medium with ampicillin (50 mg mL 1) and apramycin (50 mg mL 1) to obtain BW25141/DhitA-pUC119. The desired plasmid DhitA-pUC119 was isolated and digested with HindIII and EcoRI, and the obtained DNA fragment was ligated into the corresponding sites of pWHM3 to obtain plasmid DhitA-pW in E. coli DH5a. DhitA-pW was then introduced into E. coli ET12567/ pUZ8002 by electroporation to obtain E. coli ET12567/pUZ8002/ DhitA-pW, which was cultured in LB with apramycin (50 mg mL 1), kanamycin (50 mg mL 1), and chloramphenicol (25 mg mL 1) at 30 8C overnight. Pre-culture (1 mL) was inoculated into LB (100 mL) and incubated (30 8C, 200 rpm) until the OD590 reached 0.13. The culture was centrifuged (2200 g, 10 min), and the pelleted cells were washed twice with LB and then suspended in LB (5 mL). An aliquot (10 mL) of S. scabrisporus spore suspension was prepared according to a standard method;[11] this was mixed with the E. coli ET12567/ pUZ8002/DhitA-pW suspension (300 mL) and then spread on an ISP4 (1 % soluble starch, 0.1 % K2HPO4, 0.1 % MgSO4·7 H2O, 0.1 % NaCl, 0.2 % (NH4)2SO4, 0.2 % CaCO3, 0.1 mL (per 100 mL medium) of trace salts solution (0.1 % FeSO4·7 H2O, 0.1 % MnCl2·4 H2O, 0.1 % ZnSO4·7 H2O) agar plate containing MgCl2 (40 mm). The conjugation was conducted at 28 8C for 19 h. A soft agar solution (3 mL, Difco Nutrient Broth powder (0.8 %), Difco Bacto agar (0.5 %), nalidixic acid (25 mg mL 1), apramycin (50 mg mL 1)) was then poured onto the plate, which was incubated at 28 8C for 5 days. Several colonies were inoculated onto an ISP2 (1 % malt extract, 0.4 % extract yeast dried, 0.4 % glucose) agar plate that contained nalidixic acid (25 mg mL 1) and apramycin (10 mg mL 1), and incubated at 28 8C for 11 days. Several colonies were then re-inoculated onto a modified R5 agar plate [10.3 % sucrose, 0.01 % casamino acid, 1.0 % glucose, 0.5 % yeast extract, 0.025 % K2SO4, 1.0 % MgCl2·6 H2O, 0.54 % TES, 0.2 mL (per 100 mL medium) of trace element solution (0.004 % ZnCl2, 0.02 % FeCl3·6 H2O, 0.001 % CuCl2·2 H2O, 0.001 % MnCl2·4 H2O, 0.001 % Na2B4O7·10 H2O, 0.001 % (NH4)6Mo7O24·4 H2O), which was then mixed with 1 mL 0.5 % KH2PO4, 0.4 mL 5 m CaCl2, 1.5 mL 20 % glutamic acid, 0.7 mL 1 m NaOH, 1.5 mL 2 % NaNO3 (per 100 mL medium)] that contained thiostrepton (10 mg mL 1) or apramycin (50 mg mL 1), and cultured at 28 8C. The colonies were re-inoculated onto a new modified R5 agar plate containing of thiostrepton (10 mg mL 1) or apramycin (50 mg mL 1), and cultured for 7 days. This process was repeated three times. Several colonies that showed the expected phenotype (sensitive to thiostrepton and resistant to apramycin) were obtained. The genotype of the obtained DhitA strain was confirmed by PCR (expected product size: 1605 bp (wild-type), and 1426 bp (DhitA)) with primers hitA_ outer(f): 5’-CGCCC CGAAT CCAGT CGATT TCAAG-3’ and hitA_ outer(r): 5’-GACCG CTCCG CATTC GCCGA AGTAC-3’. Chromosomal DNA isolated from the putative mutant strains was used as the template. PCRs were performed with EmeraldAmp PCR Master Mix (Takara): 30 cycles of 98 8C for 10 s, 60 8C for 30 s, and 72 8C for 180 s. In addition, Southern hybridization with the probe hitM2 (1121 bp: amplified using hitM2_outer_(f): 5’-GCGTT TAGGG GTACG CGGCC CTGGC-3’ and hitM2_outer_(r): 5’-GCTGC GGCAG GCGAC
Figure 2. Enzymatic reaction with HitA (phenylalanine-2,3-aminomutase). Reaction conditions: HitA (0.1 mm) and amino acids (4 mm) in sodium phosphate (50 mm, pH 8.5), 28 8C, 30 min. The detection wavelength was 210 nm for reactions with A) l-Phe, B) d-Phe, C) (S)-b-Phe, and D) (R)-b-Phe.
tion with hit genes as probes. The cosmid containing hitA (cosmid 31) was digested with KpnI, and a 9.0-kb DNA fragment containing hitA was isolated and cloned into the corresponding site of the LITMUS 28 vector, thereby generating plasmid hitA-LT, which was transformed into E. coli DH5a. After confirming the sequence of the clone, hitA-LT was digested with SphI and XbaI, and a 5.2-kb DNA fragment containing hitA was isolated and cloned into the corresponding sites of pUC119, to generate plasmid hitA-pUC119. This plasmid was maintained in E. coli DH5a. E. coli strain BW25141/pKD78 (E. coli Genetic Stock Center (CGSG), Yale University) was transformed with hitA-pUC119, and the desired transformant (BW25141/pKD78/hitA-pUC119) was cultivated in lysogeny broth (LB) with ampicillin (50 mg mL 1) and chloramphenicol (10 mg mL 1). Oligonucleotides hitA_redirect(f) (5’-GGCGC CTTCG CGCCC CGAAT CCAGT CGATT TCAAG AATGA TTCCG GGGAT CCGTC GACC-3’) and hitA_redirect(r) (5’-CGAGA GAGCG GGACC GCTCC GCATT CGCCGA AGTAC TCATG TAGGC TGGAG CTGCT TC-3’; underline indicates complementary sequence to amplify the aac(3)IV-oriT
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
These are not the final page numbers! ÞÞ
5
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Communications 1 A cm 1 = 1 mg mL 1) was estimated by using a Nanodrop spectrophotometer (Thermo Scientific).
GAACA CATGG-3’) was performed to confirm insertion of the aac(3)IV-oriT cassette in the expected region. The obtained DhitA strain was stored as a spore stock solution at 30 8C.
HitA enzyme reactions were conducted in buffer B (50 mL) containing L-Phe, d-Phe, (S)-b-Phe, or (R)-b-Phe (4 mm), and HitA (0.1 mm) at 28 8C for 30 min. Each reaction was quenched at 100 8C for 10 min. The solution was filtered through a Millex-LG filter (0.2 mm, 4 mm; Merck Millipore), and an aliquot (5 mL) was injected into an Elite LaChrom HPLC apparatus (Hitachi system; L-2130 pump, L2455 diode array detector, detection at 210 nm) equipped with a Luna phenyl-hexyl column (3 mm, 4.6 150 mm, flow rate 0.3 mL min 1, RT; Phenomenex). A mixture (75:25) of H2O/TFA (0.05 %, v/v; solvent A) and methanol/TFA (0.05 %, v/v; solvent B) was used for elution. The stereochemistry of the HitA reaction product was determined after treatment with Marfey’s reagent (TCI America, Portland, OR), by comparison with authentic samples. An aliquot (10 mL) of the enzymatic solution was treated with Marfey’s reagent (10 mL; 1 % (w/v) in acetone) and NaHCO3 (2 mL of 1 m) at 37 8C for 1 h. The reaction was quenched by adding HCl (2 mL of 2 m) and extracted twice with ethyl acetate (500 mL). The organic layers were combined, and the solvent was removed with a vacuum evaporator. The resultant residue was dissolved in CH3OH (100 mL), and an aliquot (5 mL) was injected into the HPLC apparatus (flow rate 0.3 mL min 1, RT, detection 340 nm). A 30:70 mixture of solvent A and solvent B was used.
Culture of the DhitA strain: A colony of the DhitA strain was inoculated into of modified TSB medium (100 mL; 0.5 , tryptone (0.85 %, w/v), soytone (0.15 %, w/v), d-glucose (0.125 %, w/v), NaCl (0.25 %, w/v), KH2PO4 (0.125 %, w/v)) in a 500 mL baffled flask and cultured at 28 8C, 200 rpm, for 3 days. Next, BPS medium (100 mL; 0.3 , oatmeal (0.9 %, w/v), malt extract (0.15 %, w/v), yeast extract (0.09 %, w/v), MgSO4·7 H2O (0.012 %, w/v), CaCO3 (0.15 %, w/v), NaCl (0.03 %, w/v), soluble starch (0.9 %, w/v), MOPS (1.1 %, w/v), d-glucose (1 %, w/v) (autoclaved separately) in a 500 mL baffled flask was inoculated with pre-culture (1 mL). Next, (S)-b-Phe (19.1 mg; Peptech AD381-1, Bedford, MA) or (R)-b-Phe (19.2 mg, Peptech AL380-1) were dissolved in water and added, and the flasks were incubated (28 8C, 200 rpm, 6 days). The culture was then centrifuged (12 000 g, 40 min, 4 8C), and the pellet was treated with acetone (300 mL) in a 500 mL Erlenmeyer flask and stirred overnight. Acetone solution was decanted, and the solvent was removed in a rotary evaporator. The residual solution was extracted with ethyl acetate (3 400 mL). The combined organic layer was washed with brine and then dried over anhydrous Na2SO4. Following removal of the solvent, the crude extracts were obtained: 42 mg from 200 mL culture with (S)-b-Phe; 25 mg from 200 mL culture with (R)-b-Phe; 206 mg from 1.5 L culture without supplements. HPLC (Figure 1) and 1H NMR (Figure S3) analyses were conducted to detect hitachimycin production.
Acknowledgements This work was supported in part by Grants-in-Aid for Scientific Research on Innovative Areas from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT), Nagase Science and Technology Foundation, and Takeda Science Foundation.
HPLC was carried out using a Jasco instrument (PU-2089 pump, UV-2075 plus UV/Vis detector) equipped with a Luna phenyl-hexyl column (3 mm, 4.6 150 mm; Phenomenex). The flow rate was 0.5 mL min 1 and the detection wavelength was 300 nm (room temperature). The solvent profile of CH3CN (A) and H2O (B) was: 0– 15 min, linear gradient 40–90 % A; 15–20 min, 90 % A; 20–22 min, 90–100 % A; and 22–30 min, 100 % A.
Keywords: antibiotics · biosynthesis · genome mining · hitachimycin · macrolactam polyketide · phenylalanine-2,3aminomutase
HitA enzyme reaction analysis: hitA was amplified by PCR from chromosomal DNA with primers 5’-ACATA TGGGG TTCGA CACCG TGAC-3’ and 5’-ACTCG AGTCA GGAGA GAATT TCGCG CGC-3’ (NdeI and XhoI restriction sites underlined) and ligated into vector pMD19 (Takara). Following sequence confirmation, hitA was excised from this plasmid and inserted between the NdeI and XhoI sites of pColdI (Takara), which was then transformed into E. coli BL21(DE3). The resulting strain (hitA/pColdI/BL21(DE3)) was cultured in LB containing ampicillin (100 mg mL 1) at 37 8C with shaking (200 rpm) until the OD600 reached 0.6. The culture was cooled in an ice-water bath for 30 min, then isopropyl 1-thio-b-d-galactopyranoside (0.1 mm) was added, and culturing was continued at 15 8C overnight. The cells were harvested by centrifugation (4 700 g, 20 min), and suspended in buffer A (sodium phosphate (50 mm, pH 8.0)). The cells were disrupted with a Q55 sonicator (Qsonica, Newtown, CT) to obtain the cell-free extract. Following centrifugation (13 200 g, 20 min), the supernatant was loaded onto a His60 Ni Superflow resin column (Takara) and washed with buffer A containing imidazole (20 mm). The HitA protein was eluted with buffer A containing imidazole (200 mm) and passed through a PD-10 column (GE Healthcare) to obtain purified HitA. The pH of the HitA solution (50 mL) was adjusted to 8.5 with 1 mL of buffer B (50 mm Na-phosphate buffer, pH 8.5) using an Amicon Ultra-2 filter centrifugation device (7 000 g, 20 min; 2 mL, 10 kDa pore size; Millipore). To confirm the size of recombinant HitA (predicted size 58.1 kDa), SDSPAGE was performed (12.5 % gels). Protein concentration (A280
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
[1] I. Umezawa, H. Takeshima, K. Komiyama, Y. Koh, H. Yamamoto, M. Kawaguchi, J. Antibiot. 1981, 34, 259 – 265. [2] S. O¯mura, A. Nakagawa, K. Shibata, H. Sano, Tetrahedron Lett. 1982, 23, 4713 – 4716. [3] M. Igarashi, T. Tsuchida, N. Kinoshita, M. Kamijima, R. Sawa, T. Sawa, H. Naganawa, M. Hamada, T. Takeuchi, K. Yamazaki, M. Ishizuka, J. Antibiot. 1998, 51, 123 – 129. [4] D. Schulz, J. Nachtigall, J. Riedlinger, K. Schneider, K. Poralla, J. F. Imhoff, W. Beil, G. Nicholson, H.-P. Fiedler, R. D. Sssmuth, J. Antibiot. 2009, 62, 513 – 518. [5] M. Nakagawa, Y. Toda, K. Furihata, Y. Hayakawa, H. Seto, J. Antibiot. 1992, 45, 1133 – 1138. [6] K. Amagai, R. Takaku, F. Kudo, T. Eguchi, ChemBioChem 2013, 14, 1998 – 2006. [7] K. D. Walker, K. Klettke, T. Akiyama, R. Croteau, J. Biol. Chem. 2004, 279, 53947 – 53954. [8] Y. Shinohara, F. Kudo, T. Eguchi, J. Am. Chem. Soc. 2011, 133, 18134 – 18137. [9] B. Gust, G. L. Challis, K. Fowler, T. Kieser, K. F. Chater, Proc. Natl. Acad. Sci. USA 2003, 100, 1541 – 1546. [10] F. Kudo, A. Miyanaga, T. Eguchi, Nat. Prod. Rep. 2014, 31, 1056 – 1073. [11] T. Kieser, M. J. Bibb, M. J. Buttner, K. F. Chater, D. A. Hopwood, Practical Streptomyces Genetics, The John Innes Foundation, Norwich, 2000. Received: January 26, 2015 Published online on && &&, 0000
6
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
COMMUNICATIONS A very palpable hit: Genome scanning of a hitachimycin-producing Streptomyces scabrisporus strain uncovered a gene cluster that encodes a phenylalanine2,3-aminomutase (PAM), five polyketide synthases, five starter biosynthetic enzymes, and six modification enzymes. Involvement of the PAM in hitachimycin biosynthesis was confirmed through gene inactivation and enzymatic analysis of HitA.
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
These are not the final page numbers! ÞÞ
F. Kudo, K. Kawamura, A. Uchino, A. Miyanaga, M. Numakura, R. Takayanagi, T. Eguchi* && – && Genome Mining of the Hitachimycin Biosynthetic Gene Cluster: Involvement of a Phenylalanine-2,3aminomutase in Biosynthesis
7
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Communications
Genome Mining of the Hitachimycin Biosynthetic Gene Cluster: Involvement of a Phenylalanine-2,3-aminomutase in Biosynthesis Fumitaka Kudo,[a] Koichi Kawamura,[a] Asuka Uchino,[a] Akimasa Miyanaga,[a] Mario Numakura,[a] Ryuichi Takayanagi,[b] and Tadashi Eguchi*[b] Hitachimycin is a macrolactam antibiotic with (S)-b-phenylalanine (b-Phe) at the starter position of its polyketide skeleton. To understand the incorporation mechanism of b-Phe and the modification mechanism of the unique polyketide skeleton, the biosynthetic gene cluster for hitachimycin in Streptomyces scabrisporus was identified by genome mining. The identified gene cluster contains a putative phenylalanine-2,3-aminomutase (PAM), five polyketide synthases, four b-amino-acid-carrying enzymes, and a characteristic amidohydrolase. A hitA knockout mutant showed no hitachimycin production, but antibiotic production was restored by feeding with (S)-b-Phe. We also confirmed the enzymatic activity of the HitA PAM. The results suggest that the identified gene cluster is responsible for the biosynthesis of hitachimycin. A plausible biosynthetic pathway for hitachimycin, including a unique polyketide skeletal transformation mechanism, is proposed.
the starter position of their polyketides (Scheme 1). Thus, common modification enzymes are proposed to be responsible for this unique transformation. Although the cremimycin biosynthetic gene cluster has recently been identified and several unassigned genes are speculated to be involved in the unique polyketide modification,[6] the modification mechanism remains unknown. Thus, identification of comparable biosynthetic genes should provide insight into the unique multi-step enzymatic polyketide modification mechanism. Although the mechanism of b-Phe incorporation into the polyketide during hitachimycin synthesis is unclear, it is likely to be similar to related biosynthetic pathways for macrolactam antibiotics, including those for cremimycin, incednine, and vicenistatin. To further our knowledge of b-amino acid biosynthetic machinery and selective incorporation mechanisms, this study aimed to identify the hitachimycin biosynthetic gene cluster. This knowledge will help in the development of rational biosynthesis protocols for this class of polyketide. We used a genome-scanning approach to identify the hitachimycin biosynthetic gene cluster in the producer strain S. scabrisporus JCM 11712. Chromosomal DNA was extracted and sequenced with a combination of an Illumina HiSeq 2000 and a Roche Genome Sequencer FLX (mate pair sequencing analysis) to obtain ~ 11 Mb of draft genome sequence. In silico screening with the functionally characterized Taxus chinensis PAM (involved in taxol biosynthesis)[7] showed a contiguous gene cluster containing genes for a putative PAM, five polyketide synthases (PKSs), four characteristic b-amino acid-carrying enzymes, and one amidohydrolase (Table 1, Scheme 2). No other candidate gene cluster with these features was found in the genome sequence of JCM 11712. Thus, we named this the “hit” gene cluster (hitachimycin biosynthesis). During preparation of this manuscript, the draft genome sequence of a similar strain (S. scabrisporus DSM 41855) was deposited; this contains the same genes for hitachimycin biosynthesis as JCM 11712; however, several PKS genes, such as hitP2 and hitP3, are absent from the sequence data in the draft genome sequence of S. scabrisporus DSM 41855. Thus, our sequence analysis provides a more complete sequence of the hitachimycin biosynthetic gene cluster. Based on bioinformatic analysis of the genes, we propose a pathway for hitachimycin biosynthesis (Scheme 2). First, the putative PAM HitA converts l-Phe to (S)-b-phenylalanine because of the stereochemistry at C-21 of hitachimycin. (S)-b-Phe is recognized by a b-amino-acid-selective ATP-dependent ligase (HitB) and transferred to the discrete acyl carrier protein
Hitachimycin (also known as stubomycin) was isolated from Streptomyces scabrisporus and shown to be an antiprotozoal and antitumor antibiotic, as well as having potent antimicrobial activity against Bacillus subtilis and Micrococcus luteus.[1] However, its low solubility (in any solvent) has prevented examination of its precise mode of action. The main structural features of hitachimycin are the unique b-phenylalanine (b-Phe) starter unit of its polyketide skeleton and a bicyclic structure with an internal five-membered carbocycle (Scheme 1).[2] Incorporation studies illustrated that the bPhe starter unit is derived from a-phenylalanine, presumably by the action of phenylalanine-2,3-aminomutase (PAM).[2] To construct the five-membered carbocyclic moiety within the macrocyclic skeleton, unknown enzymes encoded in the biosynthetic gene cluster modify the elongated polyketide chain or macrocycle with the b-Phe starter unit. Similar five-membered cyclic structures are found in cremimycin,[3] piceamycin,[4] and viridenomycin,[5] which also have b-amino acids at [a] Prof. Dr. F. Kudo, K. Kawamura, A. Uchino, Dr. A. Miyanaga, M. Numakura Department of Chemistry, Tokyo Institute of Technology 2-12-1 O-okayama, Meguro-ku, Tokyo 152-8551 (Japan) [b] R. Takayanagi, Prof. Dr. T. Eguchi Department of Chemistry and Materials Science Tokyo Institute of Technology 2-12-1 O-okayama, Meguro-ku, Tokyo 152-8551 (Japan) E-mail: [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201500040.
ChemBioChem 0000, 00, 0 – 0
These are not the final page numbers! ÞÞ
1
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Communications
Scheme 1. Hitachimycin and related b-amino-acid-containing macrolactams.
(ACP), HitD, to form (S)-b-Phe-HitD. Then, another ATP-dependent ligase HitE activates a certain amino acid, which is transferred to the amino group of (S)-b-Phe-HitD to form aminoacyl-(S)-b-Phe-HitD. The aminoacyl-(S)-b-Phe moiety of dipeptidyl-HitD is transferred to a loading ACP domain of PKS HitP1 by the aminoacyltransferase HitC. This incorporation of (S)-b-Phe onto the PKS assembly line is same as in the vicenistatin biosynthetic pathway, except for the decarboxylation step.[8] Based on the domain structures and substrate specificity of the acyltransferase (AT) domain of the PKSs, HitP1, HitP2, HitP3, HitP4, and HitP5, a rational PKS assembly line was predicted (Scheme 2). Although the mechanism of formation of the C6=C7 bond with presumably Z stereochemistry cannot be predicted from the sequence, the other predictions for the polyketide chain were consistent with the chemical structure of hitachimycin (alignment of PKS domains in Figure S1 in the Supporting Information). The terminal aminoacyl group in the elongated polyketide chain is removed by the amidohydrolase HitF, followed by macrocyclization catalyzed by the thioesterase (TE) domain in the PKS, HitP3. The macrocyclic polyketide released from the PKS assembly line appeared to be modified by post-PKS modification enzymes. Interestingly, five genes (hitM1, hitM2, hitM3, hitM4, and hitM5) are conserved between hitachimycin and cremimycin biosynthetic gene clusters; this indicated that these genes are involved in forming the characteristic five-membered carbocycle of the polyketide skeleton (Figure S2).[6] Because HitM1 and HitM4 are putative short-chain dehydrogenases/reductases, and HitM3 is a putative cytochrome P450, oxidation reactions likely trigger the C C bond formation between C8 and C12 of the polyketide chain. The hydroxy group at C-11 is oxidized by either HitM1 or HitM4. It appears that HitM3 (cytochrome P450) catalyzes hydroxylation at C-10 and converts the hydroxy group to a carbonyl group. The presumptive a,b-unsaturated
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
ketone intermediate would be converted to a five-membered carbocycle by Michael addition or a Nazarov cyclization mechanism, possibly by HitM2 and HitM5 (two unassigned putative sugar-isomerase-type enzymes), which might act as acid/base catalysts. The highly oxidized intermediate is then reduced by either HitM1 or HitM4, and finally the putative methyltransferase HitM6 completes the biosynthesis of hitachimycin by methylation of the hydroxy group at C-10. The oxidation states of the biosynthetic intermediates generated by the redox enzymes are likely to be important for understanding this cyclization mechanism. The hit gene cluster also contains four putative regulatory genes (hitR1, hitR2, hitR3, and hitR4) and a putative transporter gene (hitT1). Thus, it appears that the identified hit gene cluster is responsible for the biosynthesis of hitachimycin. To confirm that this gene cluster is responsible for hitachimycin biosynthesis, we knocked out hitA by inserting an apramycin-resistance gene cassette (by using PCR-based gene replacement for Streptomyces genome manipulation).[9] The obtained DhitA mutant did not produce hitachimycin, thus indicating that this gene is involved in hitachimycin synthesis (Figure 1 A). When we supplemented the DhitA strain with (S)-bPhe, hitachimycin production was restored (Figures 1 B and S3). In contrast, supplementation with (R)-b-Phe did not lead to production of a corresponding diastereomer of hitachimycin or any related compound (Figure 1 C); presumably the stereochemistry of b-Phe is critical for recognition by the b-aminoacid-selective ATP-dependent ligase HitB or downstream biosynthetic enzymes. The stereochemistry at C-21 might be critical for macrocyclization of the polyketide chain. These results illustrate that hitA is involved in hitachimycin biosynthesis, and that the protein product is a PAM that provides (S)-b-Phe. Next, we examined the PAM enzymatic activity of HitA in vitro. Recombinant HitA was expressed in Escherichia coli and 2
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Communications Table 1. Hitachimycin biosynthesis genes and putative functions. Gene
Size [bp] (aa)
Accession numbers of encoded protein homologues Positives [%]/ Identity [%]
hitM1 hitM2 hitA hitP1
870 864 1551 11949
WP_020550023 WP_020550024 WP_026218103 WP_020550026
(290) (288) (517) (3983)
(DSM 41855),[b] (DSM 41855),[b] (DSM 41855),[b] (DSM 41855),[b]
cmiM1 (BAO66515), 71/79 cmiM2 (BAO66516),71/80 encP (AAF81735), 73/85 PKS, WP_014060784
hitB
1599 (533)
WP_026218104 (DSM 41855),[b] MlaJ (ACO94489), 61/73
hitC hitD hitE
945 (315) 246 (82) 1548 (516)
WP_020550028 (DSM 41855),[b] CmiS5, BAO66532, 60/71 WP_020550029 (DSM 41855),[b] MlaS, ACO94491, 57/70 WP_020550030 (DSM 41855),[b] MlaL, ACO94492, 64/73
hitR1
720 (240)
hitM3 hitM4 hitR2
1236 (412) 864 (288) 435 (145)
hitT1
1614 (538)
hitP2 hitF hitP3 hitP4
9975 906 6012 15558
(3325) (302) (2004) (5186)
hitP5 hitM5 hitM6
4836 (1612) 903 (301) 651 (217)
hitR3 hitR4
2652 (884) 2802 (934)
WP_020550031 (DSM 41855),[b] Streptomyces katrae, TetR family transcriptional regulator, WP_030302823, 47/63 WP_020550032 (DSM 41855),[b] cmiM4 (BAO66524), 73/84 WP_020550033 (DSM 41855),[b] cmiM3 (BAO66523), 71/81 WP_026218105 (DSM 41855),[b] cold-shock protein, (WP_018381685), 45/58 WP_026218106 (DSM 41855),[b] MFS transporter, (WP_030976952), 50/69 PKS, (WP_014060789), 56/66 WP_020549776 (DSM 41855),[b] cmiM6 (BAO66540), 75/85 cmiP6 (BAO66541), 65/75 WP_020550052 (DSM 41855),[b] cmiP7 (BAO66542), 67/76 WP_026219432 (DSM 41855),[b] cmiP8 (BAO66543), 65/75 WP_026219433,[b] cmiM7(BAO66544), 69/83 WP_020556451 (DSM 41855),[b] O-methyltransferase (EYF02720), 49/67 cmiR5 (BAO66545), 47/58 Orf5, (BAO66546), 47/60
Putative function
Homologues in vin, cmi, and idn clusters[c] (except for PKS)
short-chain dehydrogenase sugar phosphate isomerase/epimerase phenylalanine-2,3-aminomutase PKS (ACP-KS-AT-DH-KR-ACP-KS-AT-DHER-KR-ACP) ATP-dependent b-aminoacyl-ACP synthetase ACP aminoacyltransferase ACP ATP-dependent aminoacyl-ACP synthetase TetR family transcriptional regulator
cmiM1 cmiM2
cytochrome P450 monooxygenase 3-ketoacyl-ACP reductase cold-shock protein (regulator)
vinN, cmiS6, idnL1 vinK, cmiS5, idnL2 vinL, cmiS4, idnL6 vinM, cmiS3, idnL7
cmiM4 cmiM3
MFS transporter PKS (KS-AT-KR-ACP-KS-AT-DH-KR[a]-ACP) amidohydrolase PKS (KS-AT-DH-KR-ACP-TE) PKS (KS-AT-DH-KR-ACP-KS-AT-DHKR-ACP-KS-AT-DH-KR-ACP) PKS (KS-AT-KR-ACP) sugar phosphate isomerase/epimerase O-methyltransferase LuxR-family transcriptional regulator LuxR-family transcriptional regulator
vinJ, cmiM6, idnL5
cmiM7
cmiR5
[a] This KR domain appeared to be inactive because of the lack of a catalytic tyrosine and NADPH-binding motif. [b] These genes are in the draft genome sequence of S. scabrisporus DSM 41855. [c] vin: vicenistatin biosynthetic gene; cmi: cremimycin biosynthetic gene; idn: incednine biosynthetic gene. PKS: polyketide synthase, ACP: acyl carrier protein, KS: b-ketosynthase, AT: acyltransfearse, DH: dehydratase, ER: enoylreductase, KR: b-ketoreductase, TE: thioesterase.
Figure 1. Inactivation of hitA. HPLC analysis (detection at 300 nm) of the crude extracts from the indicated cultures.
indicating that HitA has a strict stereoselective PAM reaction mechanism. This enzymatic activity of HitA supports the hypothesis that the hit gene cluster is responsible for hitachimycin biosynthesis. In summary, we identified the hitachimycin biosynthesis gene cluster (hit) by genome mining, and characterized the (S)b-Phe-forming PAM, HitA. The identified hit gene cluster is
purified by affinity chromatography (see the Supporting Information). As expected, HitA specifically converted l-Phe to (S)b-Phe (Figure 2 A). The reverse reaction (from (S)-b-Phe to lPhe) was also observed (Figure 2 C). We confirmed the stereochemistry of the HitA reaction product to be the S form after 2,4-dinitrophenyl-5-l-alanine amide derivatization with Marfey’s reagent (Figure S5). Neither d-Phe nor (R)-b-Phe reacted, thus ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
These are not the final page numbers! ÞÞ
3
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Communications
Scheme 2. Hitachimycin biosynthetic gene cluster and the proposed biosynthesis pathway for hitachimycin.
Experimental Section
a novel member of the b-amino-acid-containing macrolactam biosynthesis group. The presence of both b-amino-acid-forming and b-amino-acid-carrying enzymes is critical for identification of this family of macrolactam polyketide biosynthetic gene clusters. In fact, a wide variety of natural products, such as isoprenoids and polyketides, are obtained by combinations of PAMs and other enzymes involved in metabolism.[10] The accumulation of knowledge of biosynthetic genes and enzymes is important to allow manipulation of the biosynthetic machinery to obtain new compounds. Mutasynthesis with various structurally similar b-amino acids by using the DhitA strain is now in progress. However, to achieve rationally engineered biosynthesis, the substrate-recognition mechanisms of HitB, HitC, HitE, and PKS enzymes should be elucidated before they can be altered.
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
Genome sequence analysis: S. scabrisporus JCM 11712 chromosomal DNA was extracted by standard phenol/chloroform extraction.[11] DNA sequencing and assembly of the obtained DNA sequence data were conducted by Takara. Automated gene annotation was performed with the MiGAP server (http://www.migap.org/ index.php/en), and the annotated open reading frames were revised by FramePlot analysis (http://www0.nih.go.jp/ ~ jun/cgi-bin/ frameplot.pl). In silico screening with the T. chinensis PAM gene and routine genetic analyses were conducted by using Geneious (Biomatters, Auckland, New Zealand). The 80-kb DNA sequence including the putative hitachimycin biosynthetic gene cluster was deposited in the DDBJ under the accession number LC008143. Disruption of hitA: A pOJ446-based cosmid library in Escherichia coli XL1 Blue MRF’ was constructed from chromosomal DNA. Several hit-gene-containing cosmids were selected by colony hybridiza-
4
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Communications cassette) were used to introduce an apramycin gene cassette into the middle of hitA. A DNA fragment (1384 bp) from plasmid pIJ773 (Plant Bioscience Limited, Norwich, UK) digested with HindIII and EcoRI was used as a template to prepare the hitA disruption DNA fragment. The PCR reaction mixture (10 mL) comprised PrimeSTAR GXL reaction buffer (2 mL), dNTPs (0.8 mL, 2.5 mm), primers (0.2 mL, 10 mm each), template (3 mL, 13 ng mL 1), and PrimeSTAR GXL polymerase (0.2 mL; Takara). Thermal cycling conditions were: 10 cycles of 94 8C for 45 s, 50 8C for 45 s, and 72 8C for 90 s, followed by 15 cycles of 94 8C for 45 s, 55 8C for 45 s, and 72 8C for 90 s. The amplified DNA fragment was introduced into BW25141/pKD78/ hitA-pUC119 by electroporation and the desired transformant was cultured in LB medium with ampicillin (50 mg mL 1) and apramycin (50 mg mL 1) to obtain BW25141/DhitA-pUC119. The desired plasmid DhitA-pUC119 was isolated and digested with HindIII and EcoRI, and the obtained DNA fragment was ligated into the corresponding sites of pWHM3 to obtain plasmid DhitA-pW in E. coli DH5a. DhitA-pW was then introduced into E. coli ET12567/ pUZ8002 by electroporation to obtain E. coli ET12567/pUZ8002/ DhitA-pW, which was cultured in LB with apramycin (50 mg mL 1), kanamycin (50 mg mL 1), and chloramphenicol (25 mg mL 1) at 30 8C overnight. Pre-culture (1 mL) was inoculated into LB (100 mL) and incubated (30 8C, 200 rpm) until the OD590 reached 0.13. The culture was centrifuged (2200 g, 10 min), and the pelleted cells were washed twice with LB and then suspended in LB (5 mL). An aliquot (10 mL) of S. scabrisporus spore suspension was prepared according to a standard method;[11] this was mixed with the E. coli ET12567/ pUZ8002/DhitA-pW suspension (300 mL) and then spread on an ISP4 (1 % soluble starch, 0.1 % K2HPO4, 0.1 % MgSO4·7 H2O, 0.1 % NaCl, 0.2 % (NH4)2SO4, 0.2 % CaCO3, 0.1 mL (per 100 mL medium) of trace salts solution (0.1 % FeSO4·7 H2O, 0.1 % MnCl2·4 H2O, 0.1 % ZnSO4·7 H2O) agar plate containing MgCl2 (40 mm). The conjugation was conducted at 28 8C for 19 h. A soft agar solution (3 mL, Difco Nutrient Broth powder (0.8 %), Difco Bacto agar (0.5 %), nalidixic acid (25 mg mL 1), apramycin (50 mg mL 1)) was then poured onto the plate, which was incubated at 28 8C for 5 days. Several colonies were inoculated onto an ISP2 (1 % malt extract, 0.4 % extract yeast dried, 0.4 % glucose) agar plate that contained nalidixic acid (25 mg mL 1) and apramycin (10 mg mL 1), and incubated at 28 8C for 11 days. Several colonies were then re-inoculated onto a modified R5 agar plate [10.3 % sucrose, 0.01 % casamino acid, 1.0 % glucose, 0.5 % yeast extract, 0.025 % K2SO4, 1.0 % MgCl2·6 H2O, 0.54 % TES, 0.2 mL (per 100 mL medium) of trace element solution (0.004 % ZnCl2, 0.02 % FeCl3·6 H2O, 0.001 % CuCl2·2 H2O, 0.001 % MnCl2·4 H2O, 0.001 % Na2B4O7·10 H2O, 0.001 % (NH4)6Mo7O24·4 H2O), which was then mixed with 1 mL 0.5 % KH2PO4, 0.4 mL 5 m CaCl2, 1.5 mL 20 % glutamic acid, 0.7 mL 1 m NaOH, 1.5 mL 2 % NaNO3 (per 100 mL medium)] that contained thiostrepton (10 mg mL 1) or apramycin (50 mg mL 1), and cultured at 28 8C. The colonies were re-inoculated onto a new modified R5 agar plate containing of thiostrepton (10 mg mL 1) or apramycin (50 mg mL 1), and cultured for 7 days. This process was repeated three times. Several colonies that showed the expected phenotype (sensitive to thiostrepton and resistant to apramycin) were obtained. The genotype of the obtained DhitA strain was confirmed by PCR (expected product size: 1605 bp (wild-type), and 1426 bp (DhitA)) with primers hitA_ outer(f): 5’-CGCCC CGAAT CCAGT CGATT TCAAG-3’ and hitA_ outer(r): 5’-GACCG CTCCG CATTC GCCGA AGTAC-3’. Chromosomal DNA isolated from the putative mutant strains was used as the template. PCRs were performed with EmeraldAmp PCR Master Mix (Takara): 30 cycles of 98 8C for 10 s, 60 8C for 30 s, and 72 8C for 180 s. In addition, Southern hybridization with the probe hitM2 (1121 bp: amplified using hitM2_outer_(f): 5’-GCGTT TAGGG GTACG CGGCC CTGGC-3’ and hitM2_outer_(r): 5’-GCTGC GGCAG GCGAC
Figure 2. Enzymatic reaction with HitA (phenylalanine-2,3-aminomutase). Reaction conditions: HitA (0.1 mm) and amino acids (4 mm) in sodium phosphate (50 mm, pH 8.5), 28 8C, 30 min. The detection wavelength was 210 nm for reactions with A) l-Phe, B) d-Phe, C) (S)-b-Phe, and D) (R)-b-Phe.
tion with hit genes as probes. The cosmid containing hitA (cosmid 31) was digested with KpnI, and a 9.0-kb DNA fragment containing hitA was isolated and cloned into the corresponding site of the LITMUS 28 vector, thereby generating plasmid hitA-LT, which was transformed into E. coli DH5a. After confirming the sequence of the clone, hitA-LT was digested with SphI and XbaI, and a 5.2-kb DNA fragment containing hitA was isolated and cloned into the corresponding sites of pUC119, to generate plasmid hitA-pUC119. This plasmid was maintained in E. coli DH5a. E. coli strain BW25141/pKD78 (E. coli Genetic Stock Center (CGSG), Yale University) was transformed with hitA-pUC119, and the desired transformant (BW25141/pKD78/hitA-pUC119) was cultivated in lysogeny broth (LB) with ampicillin (50 mg mL 1) and chloramphenicol (10 mg mL 1). Oligonucleotides hitA_redirect(f) (5’-GGCGC CTTCG CGCCC CGAAT CCAGT CGATT TCAAG AATGA TTCCG GGGAT CCGTC GACC-3’) and hitA_redirect(r) (5’-CGAGA GAGCG GGACC GCTCC GCATT CGCCGA AGTAC TCATG TAGGC TGGAG CTGCT TC-3’; underline indicates complementary sequence to amplify the aac(3)IV-oriT
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
These are not the final page numbers! ÞÞ
5
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Communications 1 A cm 1 = 1 mg mL 1) was estimated by using a Nanodrop spectrophotometer (Thermo Scientific).
GAACA CATGG-3’) was performed to confirm insertion of the aac(3)IV-oriT cassette in the expected region. The obtained DhitA strain was stored as a spore stock solution at 30 8C.
HitA enzyme reactions were conducted in buffer B (50 mL) containing L-Phe, d-Phe, (S)-b-Phe, or (R)-b-Phe (4 mm), and HitA (0.1 mm) at 28 8C for 30 min. Each reaction was quenched at 100 8C for 10 min. The solution was filtered through a Millex-LG filter (0.2 mm, 4 mm; Merck Millipore), and an aliquot (5 mL) was injected into an Elite LaChrom HPLC apparatus (Hitachi system; L-2130 pump, L2455 diode array detector, detection at 210 nm) equipped with a Luna phenyl-hexyl column (3 mm, 4.6 150 mm, flow rate 0.3 mL min 1, RT; Phenomenex). A mixture (75:25) of H2O/TFA (0.05 %, v/v; solvent A) and methanol/TFA (0.05 %, v/v; solvent B) was used for elution. The stereochemistry of the HitA reaction product was determined after treatment with Marfey’s reagent (TCI America, Portland, OR), by comparison with authentic samples. An aliquot (10 mL) of the enzymatic solution was treated with Marfey’s reagent (10 mL; 1 % (w/v) in acetone) and NaHCO3 (2 mL of 1 m) at 37 8C for 1 h. The reaction was quenched by adding HCl (2 mL of 2 m) and extracted twice with ethyl acetate (500 mL). The organic layers were combined, and the solvent was removed with a vacuum evaporator. The resultant residue was dissolved in CH3OH (100 mL), and an aliquot (5 mL) was injected into the HPLC apparatus (flow rate 0.3 mL min 1, RT, detection 340 nm). A 30:70 mixture of solvent A and solvent B was used.
Culture of the DhitA strain: A colony of the DhitA strain was inoculated into of modified TSB medium (100 mL; 0.5 , tryptone (0.85 %, w/v), soytone (0.15 %, w/v), d-glucose (0.125 %, w/v), NaCl (0.25 %, w/v), KH2PO4 (0.125 %, w/v)) in a 500 mL baffled flask and cultured at 28 8C, 200 rpm, for 3 days. Next, BPS medium (100 mL; 0.3 , oatmeal (0.9 %, w/v), malt extract (0.15 %, w/v), yeast extract (0.09 %, w/v), MgSO4·7 H2O (0.012 %, w/v), CaCO3 (0.15 %, w/v), NaCl (0.03 %, w/v), soluble starch (0.9 %, w/v), MOPS (1.1 %, w/v), d-glucose (1 %, w/v) (autoclaved separately) in a 500 mL baffled flask was inoculated with pre-culture (1 mL). Next, (S)-b-Phe (19.1 mg; Peptech AD381-1, Bedford, MA) or (R)-b-Phe (19.2 mg, Peptech AL380-1) were dissolved in water and added, and the flasks were incubated (28 8C, 200 rpm, 6 days). The culture was then centrifuged (12 000 g, 40 min, 4 8C), and the pellet was treated with acetone (300 mL) in a 500 mL Erlenmeyer flask and stirred overnight. Acetone solution was decanted, and the solvent was removed in a rotary evaporator. The residual solution was extracted with ethyl acetate (3 400 mL). The combined organic layer was washed with brine and then dried over anhydrous Na2SO4. Following removal of the solvent, the crude extracts were obtained: 42 mg from 200 mL culture with (S)-b-Phe; 25 mg from 200 mL culture with (R)-b-Phe; 206 mg from 1.5 L culture without supplements. HPLC (Figure 1) and 1H NMR (Figure S3) analyses were conducted to detect hitachimycin production.
Acknowledgements This work was supported in part by Grants-in-Aid for Scientific Research on Innovative Areas from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT), Nagase Science and Technology Foundation, and Takeda Science Foundation.
HPLC was carried out using a Jasco instrument (PU-2089 pump, UV-2075 plus UV/Vis detector) equipped with a Luna phenyl-hexyl column (3 mm, 4.6 150 mm; Phenomenex). The flow rate was 0.5 mL min 1 and the detection wavelength was 300 nm (room temperature). The solvent profile of CH3CN (A) and H2O (B) was: 0– 15 min, linear gradient 40–90 % A; 15–20 min, 90 % A; 20–22 min, 90–100 % A; and 22–30 min, 100 % A.
Keywords: antibiotics · biosynthesis · genome mining · hitachimycin · macrolactam polyketide · phenylalanine-2,3aminomutase
HitA enzyme reaction analysis: hitA was amplified by PCR from chromosomal DNA with primers 5’-ACATA TGGGG TTCGA CACCG TGAC-3’ and 5’-ACTCG AGTCA GGAGA GAATT TCGCG CGC-3’ (NdeI and XhoI restriction sites underlined) and ligated into vector pMD19 (Takara). Following sequence confirmation, hitA was excised from this plasmid and inserted between the NdeI and XhoI sites of pColdI (Takara), which was then transformed into E. coli BL21(DE3). The resulting strain (hitA/pColdI/BL21(DE3)) was cultured in LB containing ampicillin (100 mg mL 1) at 37 8C with shaking (200 rpm) until the OD600 reached 0.6. The culture was cooled in an ice-water bath for 30 min, then isopropyl 1-thio-b-d-galactopyranoside (0.1 mm) was added, and culturing was continued at 15 8C overnight. The cells were harvested by centrifugation (4 700 g, 20 min), and suspended in buffer A (sodium phosphate (50 mm, pH 8.0)). The cells were disrupted with a Q55 sonicator (Qsonica, Newtown, CT) to obtain the cell-free extract. Following centrifugation (13 200 g, 20 min), the supernatant was loaded onto a His60 Ni Superflow resin column (Takara) and washed with buffer A containing imidazole (20 mm). The HitA protein was eluted with buffer A containing imidazole (200 mm) and passed through a PD-10 column (GE Healthcare) to obtain purified HitA. The pH of the HitA solution (50 mL) was adjusted to 8.5 with 1 mL of buffer B (50 mm Na-phosphate buffer, pH 8.5) using an Amicon Ultra-2 filter centrifugation device (7 000 g, 20 min; 2 mL, 10 kDa pore size; Millipore). To confirm the size of recombinant HitA (predicted size 58.1 kDa), SDSPAGE was performed (12.5 % gels). Protein concentration (A280
&
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
[1] I. Umezawa, H. Takeshima, K. Komiyama, Y. Koh, H. Yamamoto, M. Kawaguchi, J. Antibiot. 1981, 34, 259 – 265. [2] S. O¯mura, A. Nakagawa, K. Shibata, H. Sano, Tetrahedron Lett. 1982, 23, 4713 – 4716. [3] M. Igarashi, T. Tsuchida, N. Kinoshita, M. Kamijima, R. Sawa, T. Sawa, H. Naganawa, M. Hamada, T. Takeuchi, K. Yamazaki, M. Ishizuka, J. Antibiot. 1998, 51, 123 – 129. [4] D. Schulz, J. Nachtigall, J. Riedlinger, K. Schneider, K. Poralla, J. F. Imhoff, W. Beil, G. Nicholson, H.-P. Fiedler, R. D. Sssmuth, J. Antibiot. 2009, 62, 513 – 518. [5] M. Nakagawa, Y. Toda, K. Furihata, Y. Hayakawa, H. Seto, J. Antibiot. 1992, 45, 1133 – 1138. [6] K. Amagai, R. Takaku, F. Kudo, T. Eguchi, ChemBioChem 2013, 14, 1998 – 2006. [7] K. D. Walker, K. Klettke, T. Akiyama, R. Croteau, J. Biol. Chem. 2004, 279, 53947 – 53954. [8] Y. Shinohara, F. Kudo, T. Eguchi, J. Am. Chem. Soc. 2011, 133, 18134 – 18137. [9] B. Gust, G. L. Challis, K. Fowler, T. Kieser, K. F. Chater, Proc. Natl. Acad. Sci. USA 2003, 100, 1541 – 1546. [10] F. Kudo, A. Miyanaga, T. Eguchi, Nat. Prod. Rep. 2014, 31, 1056 – 1073. [11] T. Kieser, M. J. Bibb, M. J. Buttner, K. F. Chater, D. A. Hopwood, Practical Streptomyces Genetics, The John Innes Foundation, Norwich, 2000. Received: January 26, 2015 Published online on && &&, 0000
6
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
COMMUNICATIONS A very palpable hit: Genome scanning of a hitachimycin-producing Streptomyces scabrisporus strain uncovered a gene cluster that encodes a phenylalanine2,3-aminomutase (PAM), five polyketide synthases, five starter biosynthetic enzymes, and six modification enzymes. Involvement of the PAM in hitachimycin biosynthesis was confirmed through gene inactivation and enzymatic analysis of HitA.
ChemBioChem 0000, 00, 0 – 0
www.chembiochem.org
These are not the final page numbers! ÞÞ
F. Kudo, K. Kawamura, A. Uchino, A. Miyanaga, M. Numakura, R. Takayanagi, T. Eguchi* && – && Genome Mining of the Hitachimycin Biosynthetic Gene Cluster: Involvement of a Phenylalanine-2,3aminomutase in Biosynthesis
7
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
