New milbemycins from mutant Streptomyces bingchenggensis X-4

Streptomyces bingchenggensis produced milbemycins, including α₁ (A₃), α₃ (A₄), β₁₃, β₁₄, α₂₈, α₂₉, α₃₀ and ST 906, four seco-milbemycins A, B, C and D, and two cyclic pentapeptides.^{1, 2, 3, 4, 5} Milbemycins belong to a 16-membered macrolide antibiotic with an outstanding activity against various kinds of mites.⁶ During a screening program for high production of A₃ and A₄, a mutant S. bingchenggensis X-4 was obtained by UV treated, N-methyl-N′-nitroso-N-nitroso-guanidine mutation and genetic manipulative techniques. Significant differences of phenotype, such as the morphology of aerial mycelia, and the metabolite HPLC profiles were observed between the wild-type S. bingchenggensis and its mutant strain X-4. In the course of investigating the metabolites of this mutant strain, three new interesting compounds milbemycin β₁₅ (1), seco-milbemycins E (2) and F (3) were isolated from the fermentation broth of S. bingchenggensis X-4. The structure of compound 1 was similar with milbemycin D, which is a highly selective and potent nematocide and insecticide.⁷ So the bioactivity of compound 1 should be further investigated. Compared with the seco-milbemycins isolated previously,^{2, 3} the hydroxy groups at C-5 were absent in compounds 2 and 3. Furthermore, all the milbemycins^{3, 8} and avermectins^{9, 10} obtained from microorganisms contain the hydroxyl at C-5, and the 5-dehydroxyl derivatives of milbemycins and avermectins can only be obtained using the synthetic methods.¹¹ So seco-milbemycins E and F may have an important role in understanding and perfecting the proposed biosynthesis pathways of milbemycins.

The producing strain S. bingchenggensis X-4 was maintained on an 1/2YM slant agar consisting of sucrose 0.4%, skim milk 0.1%, yeast extract (OXOID Basingstoke, Hampshire, UK) 0.2%, malt extract (BD Biosciences, San Jose, CA, USA) 0.5%, agar (BD Biosciences) 2.0% at 28 °C for 12 days.¹² A seed 15-l fermentor (FUS-15 L (A), Shanghai Guoqiang Bioengineering Equipment, Shanghai, China) containing 10 l of seed medium (sucrose 1.0%, polypepton 0.2%, K₂HPO₄ 0.05%, skim milk 0.05%)¹² was inoculated with 0.5 l of broth cultured in flask with seed medium. The flask with seed medium (2–6 × 10⁷ spores per ml) was cultured for 30 h at 28 °C on a rotary shake at 250 r.p.m. After incubation for 32 h, the seed broth (3 l) in the 15-l fermentor was transferred into the production 50-l fermentor (FUS-50 L (A), Shanghai Guoqiang Bioengineering Equipment) containing 30-l production medium (16.0% sucrose, 2.0% soybean powder, 0.5% yeast extract, 0.5% meat extract, 0.05% K₂HPO₄, 0.05% MgSO₄ 7H₂O, 0.005% FeSO₄ 7H₂O and 0.3% CaCO₃).¹² The culture temperature was 28 °C and the initial pH was 7.40 sterilized by sparging with steam at 121 °C for 30 min. The dissolved oxygen was maintained above 35% by adjusting the agitation speed. The initial aeration and agitation rate in the 15-l reactor was 1 vvm and 180 r.p.m., whereas those in the 50 l were 0.8 vvm and 150 r.p.m., respectively. In the process of fermentation, the pH was not controlled. The fermentation was performed for 10 days at 28 °C.

The fermentation broth (30 l) was filtered. The resulting cake was washed with water, and both filtrate and wash were discarded. Methanol (10 l) was used to extract the washed cake. The MeOH extract was evaporated under reduced pressure to ∼2 l at 45 °C and the resulting concentrate was extracted three times using an equal volume of EtOAc. The combined EtOAc phase was concentrated under reduced pressure to yield 30 g of oily substances. The residual oily substance was chromatographed on silica gel (Qingdao Haiyang Chemical Group, Qingdao, China; 100–200 mesh) and eluted with a petroleum ether–acetone mixture (100:0–50:50, v/v). The fractions eluted with the petroleum ether–acetone mixture (90:10, v/v) were combined and evaporated to obtain fraction I and the fractions eluted with the petroleum ether–acetone mixture (85:15, v/v) were pooled and concentrated to give fraction II. The fraction I was subjected to Sephadex LH-20 (GE Healthcare, Glies, UK) gel column eluting with MeOH to give subfraction I. The semi-preparative HPLC (Agilent 1100, Zorbax SB-C18, 5 μm, 250 × 9.4 mm i.d.; Agilent, Palo Alto, CA, USA) was applied to obtain pure compounds. The eluates were monitored using a photodiode array detector at 220 nm, and the flow rates were 1.5 ml min⁻¹ at room temperature. The subfraction I was further separated by semi-preparative HPLC using a solvent containing a CH₃OH–H₂O mixture (95:5, v/v) to obtain compounds 2 (t_R 14.5 min, 14 mg) and 3 (t_R 16.2 min, 18 mg). The fraction II was subjected to Sephadex LH-20 gel column eluting with MeOH to give subfraction II. The subfraction II was purified by the semi-preparative HPLC using a solvent containing a CH₃OH–CH₃CN–H₂O mixture (48:45:7, v/v/v) to obtain compound 1 (t_R 24.5 min, 13 mg).

Compound 1 (Figure 1) was isolated as colorless oil with UV (EtOH) λ_max (nm) (log ɛ): 242 (4.21) and [α]_D²⁵+87 (c 0.10, EtOH). Its molecular formula was established to be C₃₃H₅₀O₇ as deduced from the high-resolution electrospray ionization (HRESI)-MS m/z 581.3451 ((M+Na)⁺, calcd for C₃₃H₅₀O₇Na, 581.3449) and ¹³C NMR data (Table 1). The IR spectrum of 1 showed absorption bands assignable to the hydroxyl group (3450 cm⁻¹) and an ester carbonyl (1715 cm⁻¹). The ¹H NMR (400 MHz, CDCl₃) data (Table 1) of compound 1 exhibited one trans double bond at δ 5.50 (dd, J=14.6, 9.9 Hz) and δ 6.26 (dd, J=14.6, 11.2 Hz), two olefinic methyls at δ 1.59 (br s), 1.85 (br s), and four doublet aliphatic methyls at δ 0.79 (d, J=6.1 Hz), 0.84 (d, J=6.8 Hz), 1.01 (d, J=7.0 Hz) and 1.03 (d, J=6.8 Hz). Its ¹³C NMR and DEPT data (Table 1) showed 33 carbon resonances, including an ester carbonyl at δ 173.5 (s), a ketal at δ 97.4 (s), one oxygenated methylene at 57.8 (t), four oxygenated methines at δ 78.3 (d), 68.8 (d), 68.3 (d), 67.4 (d), one oxygenated quaternary carbon at δ 75.9 (s), in addition to five sp² methine carbons, three sp² quaternary carbons, seven aliphatic methylenes, four aliphatic methines and six methyls. The above NMR data showed that the compound 1 has the milbemycin skeleton. In the ¹H-¹H COSY spectrum (Figure 1), the correlations from δ 0.84 (H₃-32), 1.01 (H₃-33) and 1.85 (H-31) and the HMBC crossing signals (Figure 1) between δ 0.84, 1.01 and δ 28.3 (C-31) indicated the presence of an isopropyl moiety. Comparison of the ¹H and ¹³C NMR spectral data of compound 1 with those of milbemycin β₁₂¹³ revealed that compound 1 was similar to those of milbemycin β₁₂, except for the substituents at C-8 and C-25, where the methyl at C-8 and the ethyl group at C-25 in milbemycin β₁₂ were replaced by a hydroxymethyl group and isopropyl group in compound 1, respectively. The observed HMBC correlations (Figure 1) from H₃-32 (δ_H 0.84)/H₃-33 (δ_H 1.01) to C-25 (δ_C 78.3) and from H₂-27 (δ_H 4.18, 4.23) to C-7 (δ_C 75.9), C-9 (δ_C 129.8), C-8 (δ_C 139.5) confirmed the structural assignment of compound 1. Thus, the gross structure of compound 1 was established.

Figure 1

The structure and key correlations of ¹H-¹H COSY and HMBC of compound 1.

Table 1 ¹H- and ¹³C NMR data of milbemycin β₁₅ (1), secomilbemycins E (2) and F (3) (coupling constants in parenthesis)

In compound 1, the large J (14.6 Hz) of H-10 and H-11 and the NOESY crossing peaks (Figure 1) between H₂-27 and H-10, H₃-29 and H₂-16 indicated that the three double bonds at C-8 and C-9, C-10 and C-11, C-14 and C-15 were E. The NOESY correlations of H-17, H₃-30 and H-25, H-17 and H-19, H-11 and H₃-28 suggested that these protons having the same orientations as in milbemycin β₁₂.

From a biogenetic point of view, the relative stereochemistry of the chiral centers in compound 1 was assigned based on that of milbemycin E¹⁴ and milbemycin β_12. The assumption was supported by the similar optical rotation values of 1 and milbemycin E ([α]_D²⁷+157 (c 0.25, acetone)).

Compound 2 (Figure 2a) was obtained as colorless oil with UV (EtOH) λ_max (nm) (log ɛ): 288 (4.08) and [α]_D²⁵+29.6 (c 0.27, EtOH). Its molecular formula was established to be C₃₁H₄₂O₅ as deduced from the HRESI-MS m/z 517.3027 [M+Na]⁺ (calcd for C₃₁H₄₂O₅Na 517.3006) and ¹³C NMR data (Table 1). The IR spectrum of compound 2 showed absorption bands assignable to the hydroxyl group (3454 cm⁻¹) and an ester carbonyl (1724 cm⁻¹). The ¹H NMR (400 MHz, CDCl₃) data (Table 1) of compound 2 displayed one trans double bond at δ 5.90 (dd, J=14.9, 7.6 Hz) and δ 6.32 (dd, J=14.9, 11.2 Hz), one aromatic methyl at δ 2.39 (s), one vinylic methyl at δ 1.62 (br s), and three aliphatic methyls at δ 0.82 (d, J=6.5 Hz), 1.03 (d, J=6.7 Hz) and 1.09 (d, J=6.4 Hz). The signals at δ 7.38 (1H, dd, J=8.4, 1.6 Hz), 7.47 (1H, d, J=8.4 Hz) and 7.92 (1H, br s) in the ¹H NMR spectrum of compound 2 showed the presence of a 1, 2, 4-trisubstituted benzene ring moiety. Its ¹³C NMR and DEPT data (Table 1) revealed an ester carbonyl at δ 165.1 (s), a ketal at δ 97.4 (s), three oxygenated methines, one oxygenated methylene, in addition to 12 sp² carbons, 6 aliphatic methylene, 2 aliphatic methines and 5 methyls. Comparison of the ¹H and ¹³C NMR spectral data of compound 2 with those of seco-milbemycin C (compound 4, Figure 2a)³ revealed that compound 2 was almost identical to those of seco-milbemycin C, except for the substituent at C-5, where the hydroxy in seco-milbemycin C was replaced by a hydrogen atom in compound 2. The observed HMBC correlations (Figure 2b) from H-27 (δ 5.15) to C-1 (δ 165.1) and from H₃-26 (δ 2.39) to C-3 (δ 130.6), and from C-4 (δ 138.5) and C-5 (δ 134.9) further confirmed the structural assignment of compound 2. Thus, the gross structure of compound 2 was established.

Figure 2

The structures (a) of compounds 2, 3, 4 and 5, as well as the key correlations of HMBC and NOESY (b) compounds 2 and 3.

In compound 2, the NOESY correlations (Figure 2b) between H-27 and H-10 and between H-6 and H-9 indicated the Δ⁸ olefin was trans-double bond as in seco-milbemycin C. The E configuration of the Δ¹⁰ olefin of compound 2 was assigned by the large J (14.9 Hz) between H-10 and H-11. The NOESY correlation signal between H-29 and H-16 revealed that the double bond of C-14 and C-15 was E. Furthermore, the NOESY correlations (Figure 2a) of H-17, H₃-30 and H-25, H-17 and H-19, H-10 and H₃-28 showed the orientation of these protons. Thus, the structure of compound 2 was established. Biogenetically, the relative stereochemistry of the chiral centers in compound 2 was assigned based on that of seco-milbemycin C. The assumption was supported by the similarity of optical rotation values between compound 2 and secomilbemycin C ([α]_D²⁰+25 (c 0.20, EtOH))_.

Compound 3 (Figure 2a) was also isolated as colorless oil with [α]_D²⁵+32.9 (c 0.28, EtOH) and UV (EtOH) λ_max (nm) (log ɛ): 287 (4.11). The HRESI-MS gave the molecular formula C₃₂H₄₄O₅ (found: 531.3133 [M+Na]⁺, calcd: 531.3149)_. Its ¹H and ¹³C NMR data (Table 1) were very similar to those of compound 2. Comparison of the ¹H and ¹³C NMR data of compound 3 with those of compound 2, suggested that compound 3 was structurally related to compound 2. The difference between compounds 3 and 2 could be explained by the replacement of a methyl substituent in C-25 in compound 2 by an ethyl group in compound 3. The stereostructure of compound 3 was assigned as that of compound 2 by the analysis of the NMR data and NOESY correlation signals (Figure 2b). Further comparison of the NMR data with those of seco-milbemycin A (compound 5, Figure 2a)² confirmed the structure assignment of compound 3.

The acaricidal activity against adult two-spotted spider mites (Tetranychus urticae Koch) and mite eggs as well as nematocidal activity against Caenorhabditis elegans of compounds 1, 2 and 3 were evaluated according to the methods described previously.^{1, 4, 5} Although these compounds showed potent acaricidal and nematocidal activities (data not shown), the bioactivities were weaker than those of the commercial acaricide and nematocide milbemycins A₃/A₄. However, the discovery of compounds 1, 2 and 3 in the mutant S. bingchenggensis X-4 may shed new insight into the biosynthesis of milbemycin.