- Email: [email protected]
Identification of a plausible biosynthetic enzyme for the IM-2-type autoregulator in Streptomyces antibioticus
Biochimica et Biophysica Acta 1475 (2000) 329^336
www.elsevier.com/locate/bba
Identi¢cation of a plausible biosynthetic enzyme for the IM-2-type autoregulator in Streptomyces antibioticus Noriyasu Shikura, Takuya Nihira *, Yasuhiro Yamada Department of Biotechnology, Graduate School of Engineering, Osaka University, 2-1 Yamadaoka, Suita, Osaka 565-0871, Japan Received 23 February 2000; received in revised form 11 May 2000 ; accepted 12 May 2000
Abstract Virginiae butanolides (VBs) and IM-2 are members of Streptomyces hormones called `butyrolactone autoregulators' which regulate the antibiotic production in Streptomyces species at nanomolar concentrations. Cell-free extract of a VB-A overproducer, Streptomyces antibioticus NF-18, is capable of catalyzing the final step of the autoregulator biosynthesis, namely, the NADPH-dependent reduction of 6-dehydroVB-A. However, physico-chemical analyses of the purified enzymatic products revealed that, in addition to the VB-type isomer [(2R,3R,6S)-enantiomer], IM-2-type isomers [(2R,3R,6R)- and (2S,3S,6S)-enantiomers] were also produced from ( þ )-6-dehydroVB-A, suggesting the existence of several 6-dehydroVB-A reductases with respective stereoselectivities. The reductase activity of the crude extracts was separated into two activity peaks, peak I (major) and peak II (minor), by DEAE-5PW HPLC. Chiral HPLC analyses demonstrated that peak I enzyme and peak II enzyme catalyzed the production of (2R,3R,6S), (2R,3R,6R) and (2S,3S,6S) isomers at ratios of 46:1:3.2 and 4.9:1:1.5, respectively, indicating clearly that S. antibioticus NF-18 possesses at least two 6-dehydroVB-A reductases: one much favored toward VB-A biosynthesis, the other with relaxed stereoselectivity capable of synthesizing both VB-type and IM-2-type autoregulators. ß 2000 Elsevier Science B.V. All rights reserved. Keywords : Butyrolactone autoregulator ; Biosynthetic enzyme; IM-2; 6-Dehydrovirginiae butanolide A; Streptomyces
1. Introduction Streptomycetes are Gram-positive ¢lamentous bacteria well known both for the ability to produce numerous kinds of commercial antibiotics and for their complex morphological di¡erentiation on solid media. The antibiotic production and/or the morphological di¡erentiation are controlled in some Streptomyces species by low-molecular-weight compounds called butyrolactone autoregulators [1,2]. Their e¡ectiveness at extremely low concentrations as well as strict ligand speci¢city toward the corresponding producer strains implies that they should be regarded as Streptomyces hormones [3^9]. To date, 10 butyrolactone autoregulators have been isolated and their structures have been elucidated chemically. They share a characteristic 2,3-disubstituted Q-butyrolactone skeleton, but show minor structural di¡erences in the C-2 side chain, enabling the classi¢cation of these autoregulators
* Corresponding author. Fax: +81-6-6879-7432; E-mail : [email protected]
into three types (Fig. 1A): (i) the virginiae butanolide (VB) type possesses a 6-K-hydroxy group, as exempli¢ed by VB-A^E of Streptomyces virginiae [4,8,10,11], which controls virginiamycin production; (ii) the IM-2 type possesses a 6-L-hydroxy group, as exempli¢ed by IM-2 of Streptomyces lavendulae FRI-5 [12,13], which controls the production of a blue pigment and nucleoside antibiotics; and (iii) the A-factor type possesses a 6-keto group, as exempli¢ed by A-factor of Streptomyces griseus [1,5,14], which controls both streptomycin production and aerial mycelium formation. Using a VB-A hyperproducing strain, S. antibioticus NF-18, which produces about 1000 times more VB-A than S. virginiae [15], we have established a plausible biosynthetic pathway of butyrolactone autoregulators (Fig. 1B) [16,17] : After L-ketoacyl-CoA is formed via presumable condensation between an isovaleryl-CoA and two malonyl-CoAs in a process similar to that of polyketide biosynthesis, the L-ketoacyl-CoA couples with a dihydroxyacetone-type C3 unit derived from glycerol to create a L-keto ester, followed by intramolecular aldol condensation to form a Q-butyrolactone skeleton. Successive dehydration and reduction should lead to 6-dehydroVB-A
0304-4165 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 0 0 ) 0 0 0 8 5 - 4
BBAGEN 25045 13-7-00
330
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
(A-factor-type autoregulator). Finally, reduction of the 6-carbonyl group of the 6-dehydroVB-A will result either in VB-type autoregulators having (6S) absolute con¢guration or IM-2-type autoregulators having (6R) absolute con¢guration. Because natural VB-A and IM-2 have (2R,3R) absolute con¢guration, (2R,3R)-6-dehydroVB-A is the precursor for VB-A or IM-2. However, it remains to be elucidated whether (2S,3S) enantiomers of VB-A or IM-2 will be produced when (2S,3S)-6-dehydroVB-A exists at su¤ciently high concentrations. During the study on 6-dehydroVB-A reductase in the crude cell-free extracts of S. antibioticus NF-18 [18], we noticed that, concomitant with VB-A, an unknown compound X was enzymatically produced from 6-dehydroVBA. Isolation and physico-chemical analyses revealed that compound X is an IM-2-type isomer of VB-A, which raised the question whether the VB-type and IM-2-type autoregulators are produced by di¡erent reductases having de¢nite stereospeci¢city, or by one reductase with relaxed stereoselectivity. The 6-dehydroVB-A reductase activity was separated into two activity peaks by DEAE-5PW HPLC, and the stereoselectivity of each enzyme was determined thoroughly by isolation of the reaction products and physico-chemical analyses including chiral HPLC. 2. Materials and methods 2.1. Bacterial strain and culture conditions S. antibioticus NF-18 was used throughout this study. It was grown at 28³C for 72 h as described previously [15^18]. 2.2. Chemicals All the chemicals were of reagent or high-performance liquid chromatography (HPLC) grade, obtained from either Nacalai Tesque, Inc. (Osaka, Japan) or Wako Pure Chemical Industrial, Ltd. (Osaka, Japan). L-NADPH and L-NADH were purchased from Oriental Yeast Co., Ltd. ( þ )-6-DehydroVB-A, racemic VB-A, racemic VB-D, and racemic IM-2-type isomer of VB-A (VB-AIMÿ2 ) were chemically synthesized as described elsewhere [11,16,17]. 2.3. Preparation of crude cell-free extract All steps were performed at 4³C. Dialyzed crude cellfree extract was prepared with bu¡er A (0.02 M triethanolamine^HCl, pH 7.0, containing 20% (w/v) glycerol, 0.5 M NaCl, 5 mM 2-mercaptoethanol, 0.1 mM dithiothreitol, 0.1 mM p-(amidinophenyl)methanesulfonyl £uoride hydrochloride, 10 WM leupeptin, and 500 nM pepstatin A) as described previously [18]. Protein concentration was determined by a dye binding assay (Bio-Rad protein assay kit) using bovine serum albumin as the standard.
2.4. Preparative isolation of the enzymatic product Dialyzed enzyme solution (700 mg) was incubated in a total volume of 140 ml for 96 h under conditions similar to those for the 6-dehydroVB-A reductase assay [18] except that the reaction mixture contained 10 mM NADPH instead of 5 mM for the routine assay. Racemic VB-D as an internal standard was omitted. Enzymatic products were isolated as follows : the reaction mixture was adjusted to pH 3 with concentrated HCl, and the solution was centrifuged (16 000Ug, 20 min) to remove insoluble materials. The clari¢ed solution was applied to an active charcoal column (bed volume of 300 ml, Wako Pure Chemical Ind.) preequilibrated with water. After washing the resin with 900 ml each of water and 25% (v/v) MeOH, the retained material was eluted with 3 l of 100% MeOH, and the 100% MeOH fractions were evaporated to dryness. The residue was redissolved in water (2000 ml) and divided into 200 portions. Each portion was applied to a SEP-PAK C18 cartridge (1 ml type, Waters) preequilibrated with water. The cartridge was washed with 3 ml each of water, 10% (v/v) CH3 CN, and, lastly, with 20 ml of 50% (v/v) CH3 CN. All of the 50% CH3 CN fractions from the 200 samples were mixed, evaporated to dryness, then the residue was redissolved in 100% CH3 CN and divided into 170 portions to be lyophilized separately. Each lyophilized sample was benzoylated with benzoyl cyanide and tri-n-butylamine as described previously [18]. The benzoylation mixture was applied to a SEP-PAK C18 cartridge, and the cartridge was washed with 3 ml each of water, 50% (v/v) CH3 CN, and, lastly, with 20 ml of 100% (v/v) CH3 CN. All of the 100% CH3 CN fractions were mixed and evaporated to dryness. VB-A dibenzoate (Bz2 -VB-A) and VB-AIMÿ2 dibenzoate (Bz2 -VB-AIMÿ2 ) in the residue were puri¢ed at least twice by C18 reversephase HPLC (column, Cosmosil 5C18-AR, 10U250 mm; Nacalai Tesque, Kyoto, Japan) with 60% CH3 CN in water as the mobile phase at room temperature and at a £ow rate of 3 ml/min with detection at 230 nm. Bz2 -VB-A and Bz2 -VB-AIMÿ2 were eluted at 56.1 min and 59.9 min, respectively. Racemic standards of Bz2 -VB-A and Bz2 -VBAIMÿ2 were similarly prepared by benzoylation of racemic VB-A and VB-AIMÿ2 , respectively. Their respective retention times on C18 HPLC were 23 min and 25 min using 70% CH3 CN as solvent. Their respective identities were con¢rmed by chemical ionization mass spectrometry (CIMS) and 600 MHz 1 H-NMR spectrometry. 2.5. Chiral HPLC analysis of the enzymatic products In order to determine the optical purities of the reaction products, chiral HPLC was performed (column, Chiralpak AD, 4.6U250 mm; Daicel Co. Ltd., Tokyo, Japan; solvent, hexane/2-propanol 90/10; £ow rate 1.0 ml/min ; detection at 230 nm, 22³C). The enzymatic products were identi¢ed by comparing retention times to those of authen-
BBAGEN 25045 13-7-00
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
331
Fig. 1. (A) Three types of Q-butyrolactone autoregulators from Streptomyces species, and (B) plausible biosynthetic pathway. (A) Absolute con¢gurations of A-factor [1,14], VB-A [4,8,10,11], and IM-2 [7,12,13] have been assigned to (3R), (2R,3R,6S), and (2R,3R,6R), respectively, as depicted. (B) The condensation occurs to form a L-ketoacyl-CoA (2) between an isovaleryl-CoA and two malonyl-CoAs in a process similar to that of polyketide biosynthesis. 2 couples with a dihydroxyacetone-type C3 unit (1) derived from glycerol to create a L-keto ester (3), followed by intramolecular aldol condensation to form a Q-butyrolactone skeleton (4). Successive dehydration and reduction should lead to 6-dehydroVB-A (5). Finally, reduction of the 6-carbonyl group of 5 will result either in (2R,3R,6S)-VB-A (6) or its 6R epimer (7). When the enantiomer of 5 is reduced, (2S,3S,6R)-8 and (2S,3S,6S)-9 will be formed.
tic chiral standards. Authentic chiral standards of VB-A dibenzoate isomers were similarly prepared by chiral HPLC. Because natural VB-A dibenzoate and IM-2 dibenzoate possess a negative Cotton e¡ect at around 230 nm and their absolute con¢gurations had been determined unambiguously to be (2R,3R,6S) and (2R,3R,6R), respectively [11,13,16], the chemically synthesized enantiomers possessing a positive Cotton e¡ect at around 230 nm were designated (2S,3S,6R) and (2S,3S,6S), respectively. The racemic mixture of Bz2 -VB-A was separated into two enantiomers to yield (2S,3S,6R) isomer (retention time of 17.6 min, yield 81.2%) and (2R,3R,6S) isomer (retention time of 19.3 min, yield 81.2%). The racemic mixture of Bz2 -VB-AIMÿ2 was separated into two enantiomers to yield (2R,3R,6R) isomer (retention time of 15.0 min, yield 83.1%) and (2S,3S,6S) isomer (retention time of 18.5 min, yield 83.1%).
Ph), 7.57 (2 H, m, Ph), 7.43 (4 H, m, Ph), 5.54 (1 H, m, H-6), 4.45 (1 H, dd, J5a;5b = 11.4 Hz, J5a;3 = 5.1 Hz, H-5a), 4.40 (1 H, dd, J4a;4b = 9.6 Hz, J4a;3 = 7.2 Hz, H-4a), 4.38 (1 H, dd, J5b;3 = 5.1 Hz, H-5b), 4.13 (1 H, dd, J4b;3 = 6.6 Hz, H-4b), 2.97 (2 H, m, H-2 and 3), 2.05 (1 H, m, H-7a), 1.85 (1 H, m, H-7b), 1.50 (1 H, m, H-10), 1.39 (2 H, m, H-8), 1.21 (2 H, m, H-9), 0.82 (3 H, d, J = 6.6 Hz, H-11), 0.82 (3 H, d, J = 6.6 Hz, H-12). Synthetic Bz2 -VB-AIMÿ2 : identical with those of the enzymatic product. CI-MS and 600 MHz 1 H-NMR spectra of the enzymatic Bz2 -VB-A were identical with those of the synthetic Bz2 -VB-A [16]. Circular dichroism (CD) data of the dibenzoyl derivatives (in CH3 CN): enzymatic VB-A ; vO223 = 32.79; synthetic (2R,3R,6S) isomer, vO223 = 33.01; synthetic (2S,3S,6R) isomer, vO223 = +3.28; enzymatic IM-2-type compound, vO236 = +1.41; synthetic (2R,3R,6R) isomer, vO236 = 33.18; synthetic (2S,3S,6S) isomer, vO223 = +2.83.
2.6. Physico-chemical data of dibenzoyl derivatives of VB-A and IM-2-type compounds
2.7. Assay of 6-dehydroVB-A reductase
Enzymatic Bz2 -VB-AIMÿ2 : CI-MS m/z 439 (M+H) , 317, 123; 1 H-NMR N (CDCl3 , 600 MHz), 7.99 (4 H, m,
6-DehydroVB-A reductase activity was routinely assayed as described previously [18]. The amount of VBAIMÿ2 was measured by a method similar to that used
BBAGEN 25045 13-7-00
332
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
with VB-A by comparison of the HPLC peak area with a known amount of authentic compound. Bz2 -VB-A, Bz2 VB-AIMÿ2 , and Bz2 -VB-D were eluted at 21.0 min, 22.0 min, and 23.0 min, respectively. One unit of enzyme activity is de¢ned as the amount of enzyme necessary to produce 1 Wmol of VB-A or VB-AIMÿ2 per minute at 28³C. 2.8. Partial puri¢cation of 6-dehydroVB-A reductase by DEAE-5PW HPLC chromatography Cell extract of S. antibioticus NF-18 (from 3 g of wet mycelia) was dialyzed at 4³C overnight against two changes of a 100-fold volume of bu¡er B (0.02 M triethanolamine^HCl, pH 7.0, containing 20% (w/v) glycerol, 5 mM 2-mercaptoethanol, 0.1 mM dithiothreitol, 0.1 mM p-(amidinophenyl)methanesulfonyl £uoride hydrochloride, 10 WM leupeptin, and 500 nM pepstatin A) containing 0.1 M NaCl. The dialyzed solution was separated by DEAE-5PW HPLC (TSKgel DEAE-5PW, 7.5U75 mm; Tosoh, Tokyo, Japan) in aliquots of 500 Wl per run (45 mg protein). Proteins were eluted with a linear gradient of NaCl from 0.1 to 0.4 M (5 mM/min) in bu¡er B at a £ow rate of 0.5 ml/min and monitored by A280 . Fractions were collected every minute and assayed for 6-dehydroVB-A reductase. Products of the 6-dehydroVB-A reductase assay were puri¢ed by C18 reverse-phase HPLC and analyzed by chiral HPLC to determine their optical purities. For VBAIMÿ2 , chromatographic conditions were modi¢ed as follows : for C18 reverse-phase HPLC, the concentration of CH3 CN and the £ow rate were modi¢ed to 55% and 3.3 ml/min, respectively, in order to achieve complete separation from Bz2 -VB-D (an internal standard), and for the chiral HPLC, the concentration of the solvent and the £ow rate were hexane/2-propanol 98/2 and 2.5 ml/min, respectively. 2.9. Analytical methods C18 reverse-phase HPLC was performed with a Hitachi L-6200 or L-6250 intelligent pump equipped with a UV detector (Hitachi UV L-4000). Chiral HPLC was performed with a Jasco Tri Rotar-VI equipped with a UV detector (Jasco Uvidec-100-V). 1 H-NMR spectra were obtained on a Varian model Unity Inova600 spectrometer at 600 MHz using CHCl3 NH 7.25 as an internal reference in CDCl3 solution. CI-MS spectra were obtained on a Jeol JMS-DX-303HF spectrometer. CD spectra were obtained on a Jasco J-725 spectropolarimeter. 3. Results and discussion 3.1. Production of IM-2-type isomer from 6-dehydroVB-A by cell-free extract of S. antibioticus NF-18 By the highly sensitive assay procedure for 6-dehy-
Fig. 2. HPLC pro¢les of (A) compound X dibenzoate and (B) racemic VB-A and IM-2-type compound dibenzoate (dibenzoyl ( þ )-6 and dibenzoyl ( þ )-7) on a C18 reverse-phase HPLC column. CH3 CN (60%) in water was used as the mobile phase. The retention times of VB-A dibenzoate and IM-2-type compound dibenzoate were 56.1 and 59.9 min, respectively.
droVB-A reductase [18], which involves benzoylation of the two hydroxyl groups of the product to increase the sensitivity 100-fold and quantitation on reverse-phase HPLC, we noticed that the assay mixture with crude cell-free extracts of S. antibioticus NF-18 that underwent prolonged incubation contained a new compound (compound X dibenzoate) eluting slightly later than VB-A dibenzoate on C18 HPLC (retention time : compound X dibenzoate, 22 min; VB-A dibenzoate, 21 min). No corresponding peak appeared when the reaction was performed in the presence of NADH instead of NADPH or in the absence of enzyme (data not shown), indicating clearly that compound X is the catalytic product from 6-dehydroVB-A. To prepare a su¤cient amount of compound X for structural elucidation, we performed a large-scale reaction using 28 mg of ( þ )-6-dehydroVB-A as substrate, and puri¢ed 1.6 mg of compound X dibenzoate and 15.9 mg of VB-A dibenzoate as described in Section 2. Although compound X dibenzoate showed the same mass (m/z = 438) as that of VB-A dibenzoate (Bz2 -VB-A), its retention time on C18 reverse-phase HPLC di¡ered from that of Bz2 -VB-A (Fig. 2), but was comparable to that of the IM-2-type isomer dibenzoate (Bz2 -VB-AIMÿ2 ). Furthermore, all of the 1 H-NMR signals agreed well between compound X dibenzoate and synthetic Bz2 -VB-AIMÿ2 (Fig. 3), particularly for the regions of 2.97 ppm and 4.40 ppm where protons on C-2, 3, 4, and 5 appear, con¢rming that compound X is an IM-2-type isomer of VB-A. In order to determine the stereoselectivity of the 6-dehydroVB-A reduction by crude extracts, both the isolated VB-A dibenzoate and the compound X dibenzoate were analyzed by chiral HPLC using four optically pure stereoisomers as standards (Fig. 4). From the elution position
BBAGEN 25045 13-7-00
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
333
Fig. 3. Comparison of 1 H-NMR (600 MHz) spectra of (A) racemic VB-A dibenzoate, (B) racemic IM-2-type compound dibenzoate, and (C) enzymatic compound X dibenzoate.
on a Chiralpak AD HPLC column, the enzymatic VB-A was found to be an optically pure (2R,3R,6S) isomer, while compound X was proved to be an enantiomeric mixture of (2R,3R,6R) and (2S,3S,6S) isomers with a ratio of 32:68.
Considering that 6-dehydroVB-A used as substrate is a racemic mixture of (2R,3R) and (2S,3S) isomers, the results described above suggest that S. antibioticus NF-18 possesses at least three NADPH-dependent enzymatic pathways for the reduction of 6-dehydroVB-A ; the ¢rst
BBAGEN 25045 13-7-00
334
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
Fig. 4. Chiral HPLC pro¢les of enzymatic and synthetic VB-A dibenzoate and IM-2-type compound dibenzoate. HPLC was performed with a Chiralpak AD column (4.6 mm ID by 25 cm) at 22³C with hexane and isopropanol (90:10) as the mobile phase at a £ow rate of 1.0 ml/min. (A) VB-A. (1) Enzymatic product, (2) synthetic (2S,3S,6R) enantiomer, and (3) synthetic (2R,3R,6S) enantiomer. (B) IM-2-type compound. (1) Enzymatic product, (2) synthetic (2R,3R,6R) enantiomer, and (3) synthetic (2S,3S,6S) enantiomer.
two utilize (2R,3R) substrate to form either (2R,3R,6S)VB-A or the (2R,3R,6R)-IM-2-type compound, and the third utilizes (2S,3S) substrate to form the (2S,3S,6S)IM-2-type compound. Although we cannot exclude the minor possibility that S. antibioticus NF-18 might have another enzymatic pathway to form a very tiny amount of (2S,3S,6R)-VB-A, it is very unlikely because no corresponding product was detected even with the extremely high sensitivity (detection limit of 3 pmol) of our assay procedure. We con¢rmed that no interconversion between VB-type and IM-2-type compounds took place under our reaction conditions.
3.2. Partial puri¢cation of 6-dehydroVB-A reductase and characterization of stereoselectivity To characterize in more detail the biosynthetic enzyme(s) catalyzing the 6-dehydroVB-A reduction, the crude extracts were separated by anion exchange chromatography on a DEAE-5PW HPLC column (Fig. 5). Typically, enzymes catalyzing the reduction of 6-dehydroVB-A were eluted in two separate peaks; peak I occurred around fraction 55, peak II around fraction 66. Based on the VBA synthesizing activity, peak I (fr. 49^62), consisting of 92% of total activity (Table 1A), was the major peak. By the peak I enzyme, the diastereomeric ratio between VBtype and IM-2-type was much favored (11:1) for the synthesis of VB-type product, while the peak II enzyme produced twice the amount (2:1) of VB-type and IM-2-type products, which ¢ndings indicate that S. antibioticus NF18 has at least two 6-dehydroVB-A reductases of di¡erent stereoselectivity. Because they were not separable by gel ¢ltration HPLC, these enzymes seem to be of similar moTable 1A 6-DehydroVB-A reductase activity after DEAE-5PW HPLC chromatography
Fig. 5. Anion exchange chromatography of 6-dehydroVB-A reductases on a DEAE-5PW HPLC column. Crude extract (45 mg protein) was injected. Fractions were collected and assayed for 6-dehydroVB-A reductase activity. VB-A synthesizing activity (a) and VB-AIMÿ2 synthesizing activity (b) were measured in each fraction as described in Section 2.
Enzyme I (fr. 49^62) Enzyme II (fr. 63^72)
Relative activity (%)
Diastereomeric ratioa VB-A :IM-2 type
91.9 8.1
22.7:2.06 1.97:1
a The diastereomeric ratio between VB-A and IM-2-type compounds was estimated by 6-dehydroVB-A reductase activity as described in Section 2. Enzyme reaction was performed for 4 h. The amount of IM-2-type compound produced by Enzyme II was de¢ned as 1.
BBAGEN 25045 13-7-00
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
335
Table 1B Optical purities of enzymatic products after DEAE-5PW HPLC chromatography Optical purity (%) Enzyme I (fr. 55) Enzyme II (fr. 66)
VB-A (2R,3R,6S):(2S,3S,6R)
IM-2 type (2R,3R,6R):(2S,3S,6S)
100:0 100:0
24:76 40:60
lecular weight: about 82 000 under native conditions (data not shown). To further clarify the stereoselectivity of the two reductases, the reaction products produced by the representative fraction were puri¢ed with C18 reverse-phase HPLC and subjected to chiral HPLC analyses (Table 1B). As expected, only the (2R,3R,6S) isomer was detected for both enzymes in the formation of VB-A. In the formation of the IM-2-type compound, the peak I enzyme showed a more than three-fold preference for the (2S,3S,6S) isomer over the (2R,3R,6R) isomer, while the peak II enzyme showed a 1.5-fold higher preference for the synthesis of the (2S,3S,6S) isomer. Based on these results, the stereoselectivity of the two 6-dehydroVB-A reductases was estimated to be 46:1:3.2 for the peak I reductase (designated E1), and 4.9:1:1.5 for the peak II reductase (designated E2) in the formation of (2R,3R,6S), (2R,3R,6R) and (2S,3S,6S) isomers, respectively, which ¢ndings indicate that E1 possesses de¢nite stereospeci¢city toward the synthesis of VB-A, while E2 possesses relaxed stereoselectivity resulting in almost equal amounts of VB-A and IM-2-type compounds, although there is a minor possibility that E2 may contain more than one enzyme of tighter stereospeci¢city. Based on a survey of a large number of streptomycetes, it is considered that more than 60% of Streptomyces species have the ability to produce at least one of the three types of Q-butyrolactone, and that these compounds may therefore play important roles in the biology of this organism. Some Streptomyces species, such as S. cellulo£avus IFO 1378, S. sclerogranulatus IFO 14301, S. tubercicus IFO 13090, and S. coelicolor A3(2), are known to produce both VB-active and IM-2-active compounds [19,20], although their chemical structures have not yet been determined. S. antibioticus NF-18 produces a very small but su¤cient amount of endogenous IM-2-active compounds during cultivation [21], which can be explained well by the in vitro data presented above. To date, all attempts to show the possible involvement of VB-A in the physiology of S. antibioticus NF-18 have failed, though the ability of the two reductases to produce the IM-2-type autoregulators in vitro may lead to identi¢cation of the real autoregulator in this strain. Acknowledgements This study was supported in part by a grant from the
`Research for the Future Program' of the Japan Society for the Promotion of Science (JSPS).
References [1] E.M. Kleiner, S.A. Pliner, V.S. Soifer, V.V. Onoprienko, T.A. Balashova, B.V. Rosynov, A.S. Khokhlov, The structure of A-factor, a bioregulator from Streptomyces griseus, Bioorg. Khim. 2 (1976) 1142^1147. [2] A.S. Khokhlov, Problems of studies of speci¢c cell autoregulators (on the example of substances produced by some actinomycetes), in: S.N. Ananchenko (Ed.), Frontiers of Bioorganic Chemistry and Molecular Biology, Pergamon Press, Oxford, 1980, pp. 201^210. [3] O. Hara, T. Beppu, Mutants blocked in streptomycin production in Streptomyces griseus, J. Antibiot. 32 (1982) 349^358. [4] U. Gra«fe, W. Schade, I. Eritt, W.F. Fleck, L. Radics, A new inducer of anthracycline biosynthesis from Streptomyces viridochromogenes, J. Antibiot. 35 (1982) 1722^1723. [5] U. Gra«fe, G. Reinhardt, W. Schade, I. Eritt, W.F. Fleck, L. Radics, Interspeci¢c inducers of cytodi¡erentiation and anthracycline biosynthesis from Streptomyces bikiniensis and S. cyaneofuscatus, Biotechnol. Lett. 5 (1983) 591^596. [6] M. Yanagimoto, G. Terui, Physiological studies on staphylomycin production (II). Formation of a substance e¡ective in inducing staphylomycin production, J. Ferment. Technol. 49 (1971) 611^ 618. [7] M. Yanagimoto, T. Enatsu, Regulation of a blue pigment production by Q-nonalactone in Streptomyces sp, J. Ferment. Technol. 61 (1983) 545^550. [8] Y. Yamada, K. Sugamura, K. Kondo, M. Yanagimoto, H. Okada, The structure of inducing factors for virginiamycin production in Streptomyces virginiae, J. Antibiot. 40 (1987) 496^504. [9] K. Hashimoto, T. Nihira, S. Sakuda, Y. Yamada, IM-2, a butyrolactone autoregulator, induces production of several nucleoside antibiotics in Streptomyces sp. FRI-5, J. Ferment. Bioeng. 73 (1992) 449^ 455. [10] K. Kondo, Y. Higuchi, S. Sakuda, T. Nihira, Y. Yamada, New virginiae butanolides from Streptomyces virginiae, J. Antibiot. 42 (1989) 1837^1876. [11] S. Sakuda, Y. Yamada, Stereochemistry of butyrolactone autoregulators from Streptomyces, Tetrahedron Lett. 32 (1991) 1817^1820. [12] K. Sato, T. Nihira, S. Sakuda, M. Yamagimoto, Y. Yamada, Isolation and structure of a new butyrolactone autoregulator from Streptomyces sp. FRI-5, J. Ferment. Bioeng. 68 (1989) 170^173. [13] K. Mizuno, S. Sakuda, T. Nihira, Y. Yamada, Enzymatic resolution of 2-acyl-3-hydroxymethyl-4-butanolide and preparation of optically active IM-2, the autoregulator from Streptomyces sp. FRI-5, Tetrahedron 50 (1994) 10849^10858. [14] K. Mori, Revision of the absolute con¢guration of A-factor, Tetrahedron 39 (1983) 3107^3109. [15] H. Ohashi, Y.H. Zeng, T. Nihira, Y. Yamada, Distribution of virginiae butanolides in antibiotic-producing actinomycetes, and identi¢cation of the inducing factor from Streptomyces antibioticus as virginiae butanolide A, J. Antibiot. 42 (1989) 1191^1195. [16] S. Sakuda, A. Higashi, S. Tanaka, T. Nihira, Y. Yamada, Biosyn-
BBAGEN 25045 13-7-00
336
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
thesis of virginiae butanolide A, a butyrolactone autoregulator from Streptomyces, J. Am. Chem. Soc. 114 (1992) 663^668. [17] S. Sakuda, S. Tanaka, K. Mizuno, T. Nihira, Y. Yamada, Biosynthetic studies on virginiae butanolide A, a butyrolactone autoregulator from Streptomyces. Part 2. Preparation of possible biosynthetic intermediates and conversion experiments in a cell-free system, J. Chem. Soc. Perkin Trans. 1 (1993) 2309^2315. [18] N. Shikura, T. Nihira, Y. Yamada, Identi¢cation and characterization of 6-dehydroVB-A reductase from Streptomyces antibioticus, FEMS Microbiol. Lett. 171 (1999) 183^189.
[19] K. Hashimoto, T. Nihira, Y. Yamada, Distribution of virginiae butanolides and IM-2 in the genus Streptomyces, J. Ferment. Bioeng. 73 (1992) 61^65. [20] M. Kawabuchi, Y. Hara, T. Nihira, Y. Yamada, Production of butyrolactone autoregulators by Streptomyces coelicolor A3(2), FEMS Microbiol. Lett. 157 (2) (1997) 81^85. [21] C. Puttikhunt, S. Okamoto, T. Nakamura, T. Nihira, Y. Yamada, Distribution in the genus Streptomyces of a homolog to nusG, a gene encoding a transcriptional antiterminator, FEMS Microbiol. Lett. 110 (1993) 243^248.
BBAGEN 25045 13-7-00
www.elsevier.com/locate/bba
Identi¢cation of a plausible biosynthetic enzyme for the IM-2-type autoregulator in Streptomyces antibioticus Noriyasu Shikura, Takuya Nihira *, Yasuhiro Yamada Department of Biotechnology, Graduate School of Engineering, Osaka University, 2-1 Yamadaoka, Suita, Osaka 565-0871, Japan Received 23 February 2000; received in revised form 11 May 2000 ; accepted 12 May 2000
Abstract Virginiae butanolides (VBs) and IM-2 are members of Streptomyces hormones called `butyrolactone autoregulators' which regulate the antibiotic production in Streptomyces species at nanomolar concentrations. Cell-free extract of a VB-A overproducer, Streptomyces antibioticus NF-18, is capable of catalyzing the final step of the autoregulator biosynthesis, namely, the NADPH-dependent reduction of 6-dehydroVB-A. However, physico-chemical analyses of the purified enzymatic products revealed that, in addition to the VB-type isomer [(2R,3R,6S)-enantiomer], IM-2-type isomers [(2R,3R,6R)- and (2S,3S,6S)-enantiomers] were also produced from ( þ )-6-dehydroVB-A, suggesting the existence of several 6-dehydroVB-A reductases with respective stereoselectivities. The reductase activity of the crude extracts was separated into two activity peaks, peak I (major) and peak II (minor), by DEAE-5PW HPLC. Chiral HPLC analyses demonstrated that peak I enzyme and peak II enzyme catalyzed the production of (2R,3R,6S), (2R,3R,6R) and (2S,3S,6S) isomers at ratios of 46:1:3.2 and 4.9:1:1.5, respectively, indicating clearly that S. antibioticus NF-18 possesses at least two 6-dehydroVB-A reductases: one much favored toward VB-A biosynthesis, the other with relaxed stereoselectivity capable of synthesizing both VB-type and IM-2-type autoregulators. ß 2000 Elsevier Science B.V. All rights reserved. Keywords : Butyrolactone autoregulator ; Biosynthetic enzyme; IM-2; 6-Dehydrovirginiae butanolide A; Streptomyces
1. Introduction Streptomycetes are Gram-positive ¢lamentous bacteria well known both for the ability to produce numerous kinds of commercial antibiotics and for their complex morphological di¡erentiation on solid media. The antibiotic production and/or the morphological di¡erentiation are controlled in some Streptomyces species by low-molecular-weight compounds called butyrolactone autoregulators [1,2]. Their e¡ectiveness at extremely low concentrations as well as strict ligand speci¢city toward the corresponding producer strains implies that they should be regarded as Streptomyces hormones [3^9]. To date, 10 butyrolactone autoregulators have been isolated and their structures have been elucidated chemically. They share a characteristic 2,3-disubstituted Q-butyrolactone skeleton, but show minor structural di¡erences in the C-2 side chain, enabling the classi¢cation of these autoregulators
* Corresponding author. Fax: +81-6-6879-7432; E-mail : [email protected]
into three types (Fig. 1A): (i) the virginiae butanolide (VB) type possesses a 6-K-hydroxy group, as exempli¢ed by VB-A^E of Streptomyces virginiae [4,8,10,11], which controls virginiamycin production; (ii) the IM-2 type possesses a 6-L-hydroxy group, as exempli¢ed by IM-2 of Streptomyces lavendulae FRI-5 [12,13], which controls the production of a blue pigment and nucleoside antibiotics; and (iii) the A-factor type possesses a 6-keto group, as exempli¢ed by A-factor of Streptomyces griseus [1,5,14], which controls both streptomycin production and aerial mycelium formation. Using a VB-A hyperproducing strain, S. antibioticus NF-18, which produces about 1000 times more VB-A than S. virginiae [15], we have established a plausible biosynthetic pathway of butyrolactone autoregulators (Fig. 1B) [16,17] : After L-ketoacyl-CoA is formed via presumable condensation between an isovaleryl-CoA and two malonyl-CoAs in a process similar to that of polyketide biosynthesis, the L-ketoacyl-CoA couples with a dihydroxyacetone-type C3 unit derived from glycerol to create a L-keto ester, followed by intramolecular aldol condensation to form a Q-butyrolactone skeleton. Successive dehydration and reduction should lead to 6-dehydroVB-A
0304-4165 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 0 0 ) 0 0 0 8 5 - 4
BBAGEN 25045 13-7-00
330
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
(A-factor-type autoregulator). Finally, reduction of the 6-carbonyl group of the 6-dehydroVB-A will result either in VB-type autoregulators having (6S) absolute con¢guration or IM-2-type autoregulators having (6R) absolute con¢guration. Because natural VB-A and IM-2 have (2R,3R) absolute con¢guration, (2R,3R)-6-dehydroVB-A is the precursor for VB-A or IM-2. However, it remains to be elucidated whether (2S,3S) enantiomers of VB-A or IM-2 will be produced when (2S,3S)-6-dehydroVB-A exists at su¤ciently high concentrations. During the study on 6-dehydroVB-A reductase in the crude cell-free extracts of S. antibioticus NF-18 [18], we noticed that, concomitant with VB-A, an unknown compound X was enzymatically produced from 6-dehydroVBA. Isolation and physico-chemical analyses revealed that compound X is an IM-2-type isomer of VB-A, which raised the question whether the VB-type and IM-2-type autoregulators are produced by di¡erent reductases having de¢nite stereospeci¢city, or by one reductase with relaxed stereoselectivity. The 6-dehydroVB-A reductase activity was separated into two activity peaks by DEAE-5PW HPLC, and the stereoselectivity of each enzyme was determined thoroughly by isolation of the reaction products and physico-chemical analyses including chiral HPLC. 2. Materials and methods 2.1. Bacterial strain and culture conditions S. antibioticus NF-18 was used throughout this study. It was grown at 28³C for 72 h as described previously [15^18]. 2.2. Chemicals All the chemicals were of reagent or high-performance liquid chromatography (HPLC) grade, obtained from either Nacalai Tesque, Inc. (Osaka, Japan) or Wako Pure Chemical Industrial, Ltd. (Osaka, Japan). L-NADPH and L-NADH were purchased from Oriental Yeast Co., Ltd. ( þ )-6-DehydroVB-A, racemic VB-A, racemic VB-D, and racemic IM-2-type isomer of VB-A (VB-AIMÿ2 ) were chemically synthesized as described elsewhere [11,16,17]. 2.3. Preparation of crude cell-free extract All steps were performed at 4³C. Dialyzed crude cellfree extract was prepared with bu¡er A (0.02 M triethanolamine^HCl, pH 7.0, containing 20% (w/v) glycerol, 0.5 M NaCl, 5 mM 2-mercaptoethanol, 0.1 mM dithiothreitol, 0.1 mM p-(amidinophenyl)methanesulfonyl £uoride hydrochloride, 10 WM leupeptin, and 500 nM pepstatin A) as described previously [18]. Protein concentration was determined by a dye binding assay (Bio-Rad protein assay kit) using bovine serum albumin as the standard.
2.4. Preparative isolation of the enzymatic product Dialyzed enzyme solution (700 mg) was incubated in a total volume of 140 ml for 96 h under conditions similar to those for the 6-dehydroVB-A reductase assay [18] except that the reaction mixture contained 10 mM NADPH instead of 5 mM for the routine assay. Racemic VB-D as an internal standard was omitted. Enzymatic products were isolated as follows : the reaction mixture was adjusted to pH 3 with concentrated HCl, and the solution was centrifuged (16 000Ug, 20 min) to remove insoluble materials. The clari¢ed solution was applied to an active charcoal column (bed volume of 300 ml, Wako Pure Chemical Ind.) preequilibrated with water. After washing the resin with 900 ml each of water and 25% (v/v) MeOH, the retained material was eluted with 3 l of 100% MeOH, and the 100% MeOH fractions were evaporated to dryness. The residue was redissolved in water (2000 ml) and divided into 200 portions. Each portion was applied to a SEP-PAK C18 cartridge (1 ml type, Waters) preequilibrated with water. The cartridge was washed with 3 ml each of water, 10% (v/v) CH3 CN, and, lastly, with 20 ml of 50% (v/v) CH3 CN. All of the 50% CH3 CN fractions from the 200 samples were mixed, evaporated to dryness, then the residue was redissolved in 100% CH3 CN and divided into 170 portions to be lyophilized separately. Each lyophilized sample was benzoylated with benzoyl cyanide and tri-n-butylamine as described previously [18]. The benzoylation mixture was applied to a SEP-PAK C18 cartridge, and the cartridge was washed with 3 ml each of water, 50% (v/v) CH3 CN, and, lastly, with 20 ml of 100% (v/v) CH3 CN. All of the 100% CH3 CN fractions were mixed and evaporated to dryness. VB-A dibenzoate (Bz2 -VB-A) and VB-AIMÿ2 dibenzoate (Bz2 -VB-AIMÿ2 ) in the residue were puri¢ed at least twice by C18 reversephase HPLC (column, Cosmosil 5C18-AR, 10U250 mm; Nacalai Tesque, Kyoto, Japan) with 60% CH3 CN in water as the mobile phase at room temperature and at a £ow rate of 3 ml/min with detection at 230 nm. Bz2 -VB-A and Bz2 -VB-AIMÿ2 were eluted at 56.1 min and 59.9 min, respectively. Racemic standards of Bz2 -VB-A and Bz2 -VBAIMÿ2 were similarly prepared by benzoylation of racemic VB-A and VB-AIMÿ2 , respectively. Their respective retention times on C18 HPLC were 23 min and 25 min using 70% CH3 CN as solvent. Their respective identities were con¢rmed by chemical ionization mass spectrometry (CIMS) and 600 MHz 1 H-NMR spectrometry. 2.5. Chiral HPLC analysis of the enzymatic products In order to determine the optical purities of the reaction products, chiral HPLC was performed (column, Chiralpak AD, 4.6U250 mm; Daicel Co. Ltd., Tokyo, Japan; solvent, hexane/2-propanol 90/10; £ow rate 1.0 ml/min ; detection at 230 nm, 22³C). The enzymatic products were identi¢ed by comparing retention times to those of authen-
BBAGEN 25045 13-7-00
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
331
Fig. 1. (A) Three types of Q-butyrolactone autoregulators from Streptomyces species, and (B) plausible biosynthetic pathway. (A) Absolute con¢gurations of A-factor [1,14], VB-A [4,8,10,11], and IM-2 [7,12,13] have been assigned to (3R), (2R,3R,6S), and (2R,3R,6R), respectively, as depicted. (B) The condensation occurs to form a L-ketoacyl-CoA (2) between an isovaleryl-CoA and two malonyl-CoAs in a process similar to that of polyketide biosynthesis. 2 couples with a dihydroxyacetone-type C3 unit (1) derived from glycerol to create a L-keto ester (3), followed by intramolecular aldol condensation to form a Q-butyrolactone skeleton (4). Successive dehydration and reduction should lead to 6-dehydroVB-A (5). Finally, reduction of the 6-carbonyl group of 5 will result either in (2R,3R,6S)-VB-A (6) or its 6R epimer (7). When the enantiomer of 5 is reduced, (2S,3S,6R)-8 and (2S,3S,6S)-9 will be formed.
tic chiral standards. Authentic chiral standards of VB-A dibenzoate isomers were similarly prepared by chiral HPLC. Because natural VB-A dibenzoate and IM-2 dibenzoate possess a negative Cotton e¡ect at around 230 nm and their absolute con¢gurations had been determined unambiguously to be (2R,3R,6S) and (2R,3R,6R), respectively [11,13,16], the chemically synthesized enantiomers possessing a positive Cotton e¡ect at around 230 nm were designated (2S,3S,6R) and (2S,3S,6S), respectively. The racemic mixture of Bz2 -VB-A was separated into two enantiomers to yield (2S,3S,6R) isomer (retention time of 17.6 min, yield 81.2%) and (2R,3R,6S) isomer (retention time of 19.3 min, yield 81.2%). The racemic mixture of Bz2 -VB-AIMÿ2 was separated into two enantiomers to yield (2R,3R,6R) isomer (retention time of 15.0 min, yield 83.1%) and (2S,3S,6S) isomer (retention time of 18.5 min, yield 83.1%).
Ph), 7.57 (2 H, m, Ph), 7.43 (4 H, m, Ph), 5.54 (1 H, m, H-6), 4.45 (1 H, dd, J5a;5b = 11.4 Hz, J5a;3 = 5.1 Hz, H-5a), 4.40 (1 H, dd, J4a;4b = 9.6 Hz, J4a;3 = 7.2 Hz, H-4a), 4.38 (1 H, dd, J5b;3 = 5.1 Hz, H-5b), 4.13 (1 H, dd, J4b;3 = 6.6 Hz, H-4b), 2.97 (2 H, m, H-2 and 3), 2.05 (1 H, m, H-7a), 1.85 (1 H, m, H-7b), 1.50 (1 H, m, H-10), 1.39 (2 H, m, H-8), 1.21 (2 H, m, H-9), 0.82 (3 H, d, J = 6.6 Hz, H-11), 0.82 (3 H, d, J = 6.6 Hz, H-12). Synthetic Bz2 -VB-AIMÿ2 : identical with those of the enzymatic product. CI-MS and 600 MHz 1 H-NMR spectra of the enzymatic Bz2 -VB-A were identical with those of the synthetic Bz2 -VB-A [16]. Circular dichroism (CD) data of the dibenzoyl derivatives (in CH3 CN): enzymatic VB-A ; vO223 = 32.79; synthetic (2R,3R,6S) isomer, vO223 = 33.01; synthetic (2S,3S,6R) isomer, vO223 = +3.28; enzymatic IM-2-type compound, vO236 = +1.41; synthetic (2R,3R,6R) isomer, vO236 = 33.18; synthetic (2S,3S,6S) isomer, vO223 = +2.83.
2.6. Physico-chemical data of dibenzoyl derivatives of VB-A and IM-2-type compounds
2.7. Assay of 6-dehydroVB-A reductase
Enzymatic Bz2 -VB-AIMÿ2 : CI-MS m/z 439 (M+H) , 317, 123; 1 H-NMR N (CDCl3 , 600 MHz), 7.99 (4 H, m,
6-DehydroVB-A reductase activity was routinely assayed as described previously [18]. The amount of VBAIMÿ2 was measured by a method similar to that used
BBAGEN 25045 13-7-00
332
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
with VB-A by comparison of the HPLC peak area with a known amount of authentic compound. Bz2 -VB-A, Bz2 VB-AIMÿ2 , and Bz2 -VB-D were eluted at 21.0 min, 22.0 min, and 23.0 min, respectively. One unit of enzyme activity is de¢ned as the amount of enzyme necessary to produce 1 Wmol of VB-A or VB-AIMÿ2 per minute at 28³C. 2.8. Partial puri¢cation of 6-dehydroVB-A reductase by DEAE-5PW HPLC chromatography Cell extract of S. antibioticus NF-18 (from 3 g of wet mycelia) was dialyzed at 4³C overnight against two changes of a 100-fold volume of bu¡er B (0.02 M triethanolamine^HCl, pH 7.0, containing 20% (w/v) glycerol, 5 mM 2-mercaptoethanol, 0.1 mM dithiothreitol, 0.1 mM p-(amidinophenyl)methanesulfonyl £uoride hydrochloride, 10 WM leupeptin, and 500 nM pepstatin A) containing 0.1 M NaCl. The dialyzed solution was separated by DEAE-5PW HPLC (TSKgel DEAE-5PW, 7.5U75 mm; Tosoh, Tokyo, Japan) in aliquots of 500 Wl per run (45 mg protein). Proteins were eluted with a linear gradient of NaCl from 0.1 to 0.4 M (5 mM/min) in bu¡er B at a £ow rate of 0.5 ml/min and monitored by A280 . Fractions were collected every minute and assayed for 6-dehydroVB-A reductase. Products of the 6-dehydroVB-A reductase assay were puri¢ed by C18 reverse-phase HPLC and analyzed by chiral HPLC to determine their optical purities. For VBAIMÿ2 , chromatographic conditions were modi¢ed as follows : for C18 reverse-phase HPLC, the concentration of CH3 CN and the £ow rate were modi¢ed to 55% and 3.3 ml/min, respectively, in order to achieve complete separation from Bz2 -VB-D (an internal standard), and for the chiral HPLC, the concentration of the solvent and the £ow rate were hexane/2-propanol 98/2 and 2.5 ml/min, respectively. 2.9. Analytical methods C18 reverse-phase HPLC was performed with a Hitachi L-6200 or L-6250 intelligent pump equipped with a UV detector (Hitachi UV L-4000). Chiral HPLC was performed with a Jasco Tri Rotar-VI equipped with a UV detector (Jasco Uvidec-100-V). 1 H-NMR spectra were obtained on a Varian model Unity Inova600 spectrometer at 600 MHz using CHCl3 NH 7.25 as an internal reference in CDCl3 solution. CI-MS spectra were obtained on a Jeol JMS-DX-303HF spectrometer. CD spectra were obtained on a Jasco J-725 spectropolarimeter. 3. Results and discussion 3.1. Production of IM-2-type isomer from 6-dehydroVB-A by cell-free extract of S. antibioticus NF-18 By the highly sensitive assay procedure for 6-dehy-
Fig. 2. HPLC pro¢les of (A) compound X dibenzoate and (B) racemic VB-A and IM-2-type compound dibenzoate (dibenzoyl ( þ )-6 and dibenzoyl ( þ )-7) on a C18 reverse-phase HPLC column. CH3 CN (60%) in water was used as the mobile phase. The retention times of VB-A dibenzoate and IM-2-type compound dibenzoate were 56.1 and 59.9 min, respectively.
droVB-A reductase [18], which involves benzoylation of the two hydroxyl groups of the product to increase the sensitivity 100-fold and quantitation on reverse-phase HPLC, we noticed that the assay mixture with crude cell-free extracts of S. antibioticus NF-18 that underwent prolonged incubation contained a new compound (compound X dibenzoate) eluting slightly later than VB-A dibenzoate on C18 HPLC (retention time : compound X dibenzoate, 22 min; VB-A dibenzoate, 21 min). No corresponding peak appeared when the reaction was performed in the presence of NADH instead of NADPH or in the absence of enzyme (data not shown), indicating clearly that compound X is the catalytic product from 6-dehydroVB-A. To prepare a su¤cient amount of compound X for structural elucidation, we performed a large-scale reaction using 28 mg of ( þ )-6-dehydroVB-A as substrate, and puri¢ed 1.6 mg of compound X dibenzoate and 15.9 mg of VB-A dibenzoate as described in Section 2. Although compound X dibenzoate showed the same mass (m/z = 438) as that of VB-A dibenzoate (Bz2 -VB-A), its retention time on C18 reverse-phase HPLC di¡ered from that of Bz2 -VB-A (Fig. 2), but was comparable to that of the IM-2-type isomer dibenzoate (Bz2 -VB-AIMÿ2 ). Furthermore, all of the 1 H-NMR signals agreed well between compound X dibenzoate and synthetic Bz2 -VB-AIMÿ2 (Fig. 3), particularly for the regions of 2.97 ppm and 4.40 ppm where protons on C-2, 3, 4, and 5 appear, con¢rming that compound X is an IM-2-type isomer of VB-A. In order to determine the stereoselectivity of the 6-dehydroVB-A reduction by crude extracts, both the isolated VB-A dibenzoate and the compound X dibenzoate were analyzed by chiral HPLC using four optically pure stereoisomers as standards (Fig. 4). From the elution position
BBAGEN 25045 13-7-00
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
333
Fig. 3. Comparison of 1 H-NMR (600 MHz) spectra of (A) racemic VB-A dibenzoate, (B) racemic IM-2-type compound dibenzoate, and (C) enzymatic compound X dibenzoate.
on a Chiralpak AD HPLC column, the enzymatic VB-A was found to be an optically pure (2R,3R,6S) isomer, while compound X was proved to be an enantiomeric mixture of (2R,3R,6R) and (2S,3S,6S) isomers with a ratio of 32:68.
Considering that 6-dehydroVB-A used as substrate is a racemic mixture of (2R,3R) and (2S,3S) isomers, the results described above suggest that S. antibioticus NF-18 possesses at least three NADPH-dependent enzymatic pathways for the reduction of 6-dehydroVB-A ; the ¢rst
BBAGEN 25045 13-7-00
334
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
Fig. 4. Chiral HPLC pro¢les of enzymatic and synthetic VB-A dibenzoate and IM-2-type compound dibenzoate. HPLC was performed with a Chiralpak AD column (4.6 mm ID by 25 cm) at 22³C with hexane and isopropanol (90:10) as the mobile phase at a £ow rate of 1.0 ml/min. (A) VB-A. (1) Enzymatic product, (2) synthetic (2S,3S,6R) enantiomer, and (3) synthetic (2R,3R,6S) enantiomer. (B) IM-2-type compound. (1) Enzymatic product, (2) synthetic (2R,3R,6R) enantiomer, and (3) synthetic (2S,3S,6S) enantiomer.
two utilize (2R,3R) substrate to form either (2R,3R,6S)VB-A or the (2R,3R,6R)-IM-2-type compound, and the third utilizes (2S,3S) substrate to form the (2S,3S,6S)IM-2-type compound. Although we cannot exclude the minor possibility that S. antibioticus NF-18 might have another enzymatic pathway to form a very tiny amount of (2S,3S,6R)-VB-A, it is very unlikely because no corresponding product was detected even with the extremely high sensitivity (detection limit of 3 pmol) of our assay procedure. We con¢rmed that no interconversion between VB-type and IM-2-type compounds took place under our reaction conditions.
3.2. Partial puri¢cation of 6-dehydroVB-A reductase and characterization of stereoselectivity To characterize in more detail the biosynthetic enzyme(s) catalyzing the 6-dehydroVB-A reduction, the crude extracts were separated by anion exchange chromatography on a DEAE-5PW HPLC column (Fig. 5). Typically, enzymes catalyzing the reduction of 6-dehydroVB-A were eluted in two separate peaks; peak I occurred around fraction 55, peak II around fraction 66. Based on the VBA synthesizing activity, peak I (fr. 49^62), consisting of 92% of total activity (Table 1A), was the major peak. By the peak I enzyme, the diastereomeric ratio between VBtype and IM-2-type was much favored (11:1) for the synthesis of VB-type product, while the peak II enzyme produced twice the amount (2:1) of VB-type and IM-2-type products, which ¢ndings indicate that S. antibioticus NF18 has at least two 6-dehydroVB-A reductases of di¡erent stereoselectivity. Because they were not separable by gel ¢ltration HPLC, these enzymes seem to be of similar moTable 1A 6-DehydroVB-A reductase activity after DEAE-5PW HPLC chromatography
Fig. 5. Anion exchange chromatography of 6-dehydroVB-A reductases on a DEAE-5PW HPLC column. Crude extract (45 mg protein) was injected. Fractions were collected and assayed for 6-dehydroVB-A reductase activity. VB-A synthesizing activity (a) and VB-AIMÿ2 synthesizing activity (b) were measured in each fraction as described in Section 2.
Enzyme I (fr. 49^62) Enzyme II (fr. 63^72)
Relative activity (%)
Diastereomeric ratioa VB-A :IM-2 type
91.9 8.1
22.7:2.06 1.97:1
a The diastereomeric ratio between VB-A and IM-2-type compounds was estimated by 6-dehydroVB-A reductase activity as described in Section 2. Enzyme reaction was performed for 4 h. The amount of IM-2-type compound produced by Enzyme II was de¢ned as 1.
BBAGEN 25045 13-7-00
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
335
Table 1B Optical purities of enzymatic products after DEAE-5PW HPLC chromatography Optical purity (%) Enzyme I (fr. 55) Enzyme II (fr. 66)
VB-A (2R,3R,6S):(2S,3S,6R)
IM-2 type (2R,3R,6R):(2S,3S,6S)
100:0 100:0
24:76 40:60
lecular weight: about 82 000 under native conditions (data not shown). To further clarify the stereoselectivity of the two reductases, the reaction products produced by the representative fraction were puri¢ed with C18 reverse-phase HPLC and subjected to chiral HPLC analyses (Table 1B). As expected, only the (2R,3R,6S) isomer was detected for both enzymes in the formation of VB-A. In the formation of the IM-2-type compound, the peak I enzyme showed a more than three-fold preference for the (2S,3S,6S) isomer over the (2R,3R,6R) isomer, while the peak II enzyme showed a 1.5-fold higher preference for the synthesis of the (2S,3S,6S) isomer. Based on these results, the stereoselectivity of the two 6-dehydroVB-A reductases was estimated to be 46:1:3.2 for the peak I reductase (designated E1), and 4.9:1:1.5 for the peak II reductase (designated E2) in the formation of (2R,3R,6S), (2R,3R,6R) and (2S,3S,6S) isomers, respectively, which ¢ndings indicate that E1 possesses de¢nite stereospeci¢city toward the synthesis of VB-A, while E2 possesses relaxed stereoselectivity resulting in almost equal amounts of VB-A and IM-2-type compounds, although there is a minor possibility that E2 may contain more than one enzyme of tighter stereospeci¢city. Based on a survey of a large number of streptomycetes, it is considered that more than 60% of Streptomyces species have the ability to produce at least one of the three types of Q-butyrolactone, and that these compounds may therefore play important roles in the biology of this organism. Some Streptomyces species, such as S. cellulo£avus IFO 1378, S. sclerogranulatus IFO 14301, S. tubercicus IFO 13090, and S. coelicolor A3(2), are known to produce both VB-active and IM-2-active compounds [19,20], although their chemical structures have not yet been determined. S. antibioticus NF-18 produces a very small but su¤cient amount of endogenous IM-2-active compounds during cultivation [21], which can be explained well by the in vitro data presented above. To date, all attempts to show the possible involvement of VB-A in the physiology of S. antibioticus NF-18 have failed, though the ability of the two reductases to produce the IM-2-type autoregulators in vitro may lead to identi¢cation of the real autoregulator in this strain. Acknowledgements This study was supported in part by a grant from the
`Research for the Future Program' of the Japan Society for the Promotion of Science (JSPS).
References [1] E.M. Kleiner, S.A. Pliner, V.S. Soifer, V.V. Onoprienko, T.A. Balashova, B.V. Rosynov, A.S. Khokhlov, The structure of A-factor, a bioregulator from Streptomyces griseus, Bioorg. Khim. 2 (1976) 1142^1147. [2] A.S. Khokhlov, Problems of studies of speci¢c cell autoregulators (on the example of substances produced by some actinomycetes), in: S.N. Ananchenko (Ed.), Frontiers of Bioorganic Chemistry and Molecular Biology, Pergamon Press, Oxford, 1980, pp. 201^210. [3] O. Hara, T. Beppu, Mutants blocked in streptomycin production in Streptomyces griseus, J. Antibiot. 32 (1982) 349^358. [4] U. Gra«fe, W. Schade, I. Eritt, W.F. Fleck, L. Radics, A new inducer of anthracycline biosynthesis from Streptomyces viridochromogenes, J. Antibiot. 35 (1982) 1722^1723. [5] U. Gra«fe, G. Reinhardt, W. Schade, I. Eritt, W.F. Fleck, L. Radics, Interspeci¢c inducers of cytodi¡erentiation and anthracycline biosynthesis from Streptomyces bikiniensis and S. cyaneofuscatus, Biotechnol. Lett. 5 (1983) 591^596. [6] M. Yanagimoto, G. Terui, Physiological studies on staphylomycin production (II). Formation of a substance e¡ective in inducing staphylomycin production, J. Ferment. Technol. 49 (1971) 611^ 618. [7] M. Yanagimoto, T. Enatsu, Regulation of a blue pigment production by Q-nonalactone in Streptomyces sp, J. Ferment. Technol. 61 (1983) 545^550. [8] Y. Yamada, K. Sugamura, K. Kondo, M. Yanagimoto, H. Okada, The structure of inducing factors for virginiamycin production in Streptomyces virginiae, J. Antibiot. 40 (1987) 496^504. [9] K. Hashimoto, T. Nihira, S. Sakuda, Y. Yamada, IM-2, a butyrolactone autoregulator, induces production of several nucleoside antibiotics in Streptomyces sp. FRI-5, J. Ferment. Bioeng. 73 (1992) 449^ 455. [10] K. Kondo, Y. Higuchi, S. Sakuda, T. Nihira, Y. Yamada, New virginiae butanolides from Streptomyces virginiae, J. Antibiot. 42 (1989) 1837^1876. [11] S. Sakuda, Y. Yamada, Stereochemistry of butyrolactone autoregulators from Streptomyces, Tetrahedron Lett. 32 (1991) 1817^1820. [12] K. Sato, T. Nihira, S. Sakuda, M. Yamagimoto, Y. Yamada, Isolation and structure of a new butyrolactone autoregulator from Streptomyces sp. FRI-5, J. Ferment. Bioeng. 68 (1989) 170^173. [13] K. Mizuno, S. Sakuda, T. Nihira, Y. Yamada, Enzymatic resolution of 2-acyl-3-hydroxymethyl-4-butanolide and preparation of optically active IM-2, the autoregulator from Streptomyces sp. FRI-5, Tetrahedron 50 (1994) 10849^10858. [14] K. Mori, Revision of the absolute con¢guration of A-factor, Tetrahedron 39 (1983) 3107^3109. [15] H. Ohashi, Y.H. Zeng, T. Nihira, Y. Yamada, Distribution of virginiae butanolides in antibiotic-producing actinomycetes, and identi¢cation of the inducing factor from Streptomyces antibioticus as virginiae butanolide A, J. Antibiot. 42 (1989) 1191^1195. [16] S. Sakuda, A. Higashi, S. Tanaka, T. Nihira, Y. Yamada, Biosyn-
BBAGEN 25045 13-7-00
336
N. Shikura et al. / Biochimica et Biophysica Acta 1475 (2000) 329^336
thesis of virginiae butanolide A, a butyrolactone autoregulator from Streptomyces, J. Am. Chem. Soc. 114 (1992) 663^668. [17] S. Sakuda, S. Tanaka, K. Mizuno, T. Nihira, Y. Yamada, Biosynthetic studies on virginiae butanolide A, a butyrolactone autoregulator from Streptomyces. Part 2. Preparation of possible biosynthetic intermediates and conversion experiments in a cell-free system, J. Chem. Soc. Perkin Trans. 1 (1993) 2309^2315. [18] N. Shikura, T. Nihira, Y. Yamada, Identi¢cation and characterization of 6-dehydroVB-A reductase from Streptomyces antibioticus, FEMS Microbiol. Lett. 171 (1999) 183^189.
[19] K. Hashimoto, T. Nihira, Y. Yamada, Distribution of virginiae butanolides and IM-2 in the genus Streptomyces, J. Ferment. Bioeng. 73 (1992) 61^65. [20] M. Kawabuchi, Y. Hara, T. Nihira, Y. Yamada, Production of butyrolactone autoregulators by Streptomyces coelicolor A3(2), FEMS Microbiol. Lett. 157 (2) (1997) 81^85. [21] C. Puttikhunt, S. Okamoto, T. Nakamura, T. Nihira, Y. Yamada, Distribution in the genus Streptomyces of a homolog to nusG, a gene encoding a transcriptional antiterminator, FEMS Microbiol. Lett. 110 (1993) 243^248.
BBAGEN 25045 13-7-00