Cytotoxic Trichothecene Macrolides Produced by ... - ACS Publications

4 days ago - i Man. ,. #. and. Ren. -. X. iang Tan. §. 2. CONT. ENT. S. Table. S. 1 . Crystallographic and experimental data for. myr. othecine D (. ...
2 downloads 0 Views 3MB Size
Download PDF
Article Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX

pubs.acs.org/jnp

Cytotoxic Trichothecene Macrolides Produced by the Endophytic Myrothecium roridum Li Shen,*,†,‡,§,⊥ Chun-Zhi Ai,¶ Yong-Chun Song,§ Feng-Wu Wang,§ Rui-Hua Jiao,§ Ai-Hua Zhang,§ Hui-Zi Man,# and Ren-Xiang Tan§ †

Institute of Translational Medicine, Medical College, Yangzhou University, Yangzhou 225001, People’s Republic of China Jiangsu Key Laboratory of Integrated Traditional Chinese and Western Medicine for Prevention and Treatment of Senile Diseases, Yangzhou University, Yangzhou 225001, People’s Republic of China § Institute of Functional Biomolecules, State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University, Nanjing 210093, People’s Republic of China ⊥ Jiangsu Co-Innovation Center for the Prevention and Control of Important Animal Infectious Diseases and Zoonoses, College of Veterinary Medicine, Yangzhou University, Yangzhou 225009, People’s Republic of China ¶ Institute for Advanced Study, Laboratory of Optoelectronic Devices and Systems of Ministry of Education and Guangdong Province, College of Optoelectronic Engineering, Shenzhen University, Shenzhen 518061, People’s Republic of China # Laboratory of Pharmaceutical Resource Discovery, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, People’s Republic of China

Downloaded by UNIV OF SOUTHERN INDIANA at 17:23:34:870 on May 22, 2019 from https://pubs.acs.org/doi/10.1021/acs.jnatprod.8b01034.

‡

S Supporting Information *

ABSTRACT: Six new macrolides named myrothecines D−G (1−4), 16-hydroxymytoxin B (5), and 14′-dehydrovertisporin (6), including four 10,13-cyclotrichothecane derivatives, in addition to 12 known compounds (7−18), were isolated from three endophytic Myrothecium roridum, IFB-E008, IFB-E009, and IFB-E012. The isolated compounds were characterized by MS, NMR, CD, and single-crystal X-ray crystallography. The isolated macrolides exhibited an antiproliferation effect against chronic myeloid leukemia K562 and colorectal carcinoma SW1116 cell lines. Compounds 1−6 were cytotoxic, with IC50 values ranging between 56 nM and 16 μM. Since slight structural changes led to obvious activity differences, the CoMFA (comparative molecular field analysis) and CoMSIA (comparative molecular similarity indices analysis) methods were then used to explore the 3D QSAR (three-dimensional quantitative structure−activity relationship) of these macrolides. The result showed that the steric, electrostatic, hydrophobic, and H-bond acceptor factors were involved in their cytotoxicity and provided an in-depth understanding of the structure− activity relationships of these metabolites.

T

nanomolar level (IC50 56 nM). A detailed three-dimensional quantitative structure−activity relationship (3D QSAR) analysis indicated that the cytotoxicity of these trichothecene macrolides is dependent on steric, electrostatic, hydrophobic, and H-bondaccepting factors. The possible biosynthetic pathways from 12,13-epoxytrichothec-9-ene macrolide to the 10,13-cyclotrichothecane derivative indicated that the endosymbiosis of microorganisms with host plants may result in substantial changes in metabolite profiles. This concept is of value in diversifying secondary metabolites during the discovery process leading to new drug leads.

richothecene fungal toxins with a macrocyclic dilactone moiety are common bioactive secondary metabolites of Myrothecium species,1−5 which have attracted attention as potential anticancer drug leads.6,7 Myrothecium roridum IFBE008 and IFB-E009 are endophytic fungi residing in Trachelospermum jasminoides (Lindl.) Lem., Apocynaceae. M. roridum IFB-E012 is an endophytic fungus associated with Artemisia annua L., Asteraceae. To characterize trichothecene macrolides from endophytic M. roridum strains, fractionation and isolation of trichothecenes from these three strains guided by the 1H NMR spectra was carried out. Six new macrolides, myrothecines D−G (1−4) (four 10,13-cyclotrichothecane derivatives and named after myrothecines A−C1 in Scheme S1), 16-hydroxymytoxin B (5), and 14′-dehydrovertisporin (6), were characterized, in addition to 12 known compounds (7− 18). The new macrolides were cytotoxic against K562 and SW1116 cell lines, with 14′-dehydrovertisporin (6) active at the © XXXX American Chemical Society and American Society of Pharmacognosy

Received: December 7, 2018

A

DOI: 10.1021/acs.jnatprod.8b01034 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

ESIMS. Detailed interpretation of 1D and 2D NMR data of 2 (Tables 1 and 2) demonstrated that it was a 13′-epimer of myrothecine C (9).1 This assumption was supported by the NOE effect of H-13′ (δH 4.39) with H-7′ (δH 1.88), which was not discerned in that of myrothecine C.1 However, the CD spectrum of 2 (Figure S3) was identical to that of myrothecine C, establishing its 2R,4R,5S,6R,9S,10S,11R,12R,6′S,12′R,13′Sconfigurations.1 Myrothecine F (3), afforded as a white, amorphous powder, was demonstrated to have a molecular formula of C29H38O10, identical with that of myrothecine A (7)1 by HR-ESIMS (569.2357 calcd for C29H38O10Na). The 1H and 13C NMR data of 3 (Tables 1 and 2) showed that it was the 12′-epimer of 7.1 This was reinforced both by the CD spectrum of 3 (Figure S3), which was different from that of 7,1 and by the upfield movement of C-2′ and C-12′ signals by 3.1 and 3.3 ppm, respectively, because of the increased γ-gauche effect from 12βOH.13 Accordingly, compound 3 possesses 2R, 4R, 5S, 6R, 9S, 10S, 11R, 12R, 6′S, 12′S-configurations. Myrothecine G (4), isolated as colorless crystals, was deduced to have a molecular formula of C31H40O11 by HR-ESIMS. The 1 H and 13C NMR data of 4 (Tables 1 and 2) were similar to those of myrothecine A (7).1 However, a set of NMR signals arising from an acetyl group at δH 2.01 as well as at δC 20.7 and 169.3 suggested a 12′-O-acetyl derivative of 7. This assumption was confirmed by the HMBC correlations of C-15′ with H-12′, which appeared at δH 4.96, shifting downfield by 0.99 ppm from that of 7.1 Compound 4 was shown to share the same absolute configuration (2R, 4R, 5S, 6R, 9S, 10S, 11R, 12R, 6′S, and 12′R) as 7 according to its CD spectrum (Figure S3), which was identical to that of myrothecine A (7).1 16-Hydroxymytoxin B (5, colorless crystals) was indicated to have a molecular formula of C29H36O10 by HR-ESIMS. The 1H and 13C NMR data of 5 (Tables 1 and 2) were close to those of mytoxin B (12).4 Further scrutiny of the 1H and 13C NMR spectra of 5 and 12 underscored the presence of the 16hydroxymethyl of 5, which resonated at δH 3.97 and δC 65.6. This proposal was confirmed by the 2D NMR experimentations (1H−1H COSY, NOESY, HMQC, and HMBC), which allowed the exact assignment of all 1H and 13C NMR signals of 5. Moreover, the absolute configuration of 5 was assigned as 2R, 4R, 5S, 6R, 11R, 12R, 6′S, and 12′R by its CD curve (Figure S3), which was nearly identical to that of mytoxin B (12).1 14′-Dehydrovertisporin (6), isolated as colorless needle crystals, was determined to have a molecular formula of C29H34O10 by HR-ESIMS. The 2D NMR spectra of 6 allowed the exact assignment of all 1H and 13C NMR signals (Tables 1 and 2). The presence of a pair of mutually coupled doublets (J = 4.1 Hz) at δH 3.03 and 2.81 ppm (assigned as H-13) and the similar 1H and 13C NMR data of 6 from C/H-2 to C/H-16 to those of mytoxin B (12) indicated the presence of the 12,13epoxytrichothec-9-ene moiety common to compounds 10 and 12.1 Moreover, the 1H and 13C NMR signals due to the diacyl residue (from C/H-1′ to C/H-14′) of 6 as well as its CD curve (Figure S3) were identical to those of myrothecine C (9), indicating its 6′S, 12′R, 13′R-configurations.1 Thus, the absolute stereochemistry of 6 was addressed to be 2R, 4R, 5S, 6R, 9S, 10S, 11R, 12R, 6′S, 12′R, and 13′R. Bioactivity and 3D QSAR of Macrolides. All trichothecene macrolides displayed in vitro cytotoxicity against human chronic myeloid leukemia K562 and colorectal carcinoma SW1116 cell lines. In particular, the new compounds 1−6 were cytotoxic, with IC50 values ranging from 56 nM to 16 μM (Table

■

RESULTS AND DISCUSSION Cultures of M. roridum strains IFB-E008, IFB-E009, and IFBE012 were extracted, and these extracts fractionated by a combination of silica gel column chromatography, ODS column chromatography, Sephadex LH-20 gel filtration, and HPLC. 1H NMR was used to track trichothecene macrolides in the extracts and subfractions thereof using the H-9′ signals around δ 6.5 (ddd, J ≈ 11, 9, 6 Hz) to detect the presence of the trichothecenes. As a result, myrothecines D−G (1−4), 16hydroxymytoxin B (5), and 14′-dehydrovertisporin (6) were obtained along with myrothecines A−C (7−9),1 12′-hydroxyroridin E (10),8 roridin E (11),4 mytoxin B (12),5 roridin A (13),9 verrucarin A (14),9 14′-hydroxymytoxin B (15),10 16hydroxyroridin E (16),10 16-hydroxyverrucarin J (17),6 and trichoverritone (18).11 Structure Elucidation of New Macrolides. Myrothecine D (1), isolated as colorless needle crystals, was ascertained to have a molecular formula of C29H38O11 by high-resolution electrospray ionization mass spectrometry (HR-ESIMS). The 1 H and 13C NMR data of 1 (Tables 1 and 2) were similar to those of myrothecine C (9).1 However, the H-13′ singlet at δ 3.62 in the 1H NMR spectrum of 9 was replaced by a pair of doublets (J = 5.0 Hz) at δ 3.75 and 5.36 in that of 1. This observation, along with a hemiketal carbon resonance at δ 105.3, showed the presence of a 2,3-dihydroxytetrahydrofuran moiety in the molecule. This assignment was substantiated by the single-crystal X-ray diffraction analysis of 1 (Figures 1, S10, and S11).12 The 12′R-configuration of 1, identical to that of myrothecine A (7), was determined from the Cotton effects at 234 (negative) and 252 nm (positive) in the circular dichroism (CD) spectrum of 1, resulting respectively from π → π* and n → π* transitions of the γ-chiral α,β-unsaturated ester chromophore.1 Relative to the chirality of C-12′, the stereochemistry of 1 was convincingly determined as 2R, 4R, 5S, 6R, 9S, 10S, 11R, 12R, 6′S, 12′R, 13′S, and 14′S from its X-ray crystal structure. Myrothecine E (2), obtained as colorless crystals, was determined to have a molecular formula of C29H36O11 by HRB

DOI: 10.1021/acs.jnatprod.8b01034 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Table 1. 1H NMR Data of Compounds 1−6 (δ in ppm, J in Hz) 1a no. 2 3 4 7 8 10 11 13 14 15 16 2′ 4′ 5′ 7′ 8′ 9′ 10 12′ 13′ 14′ 16′

δH

2b

3c

4d

5b

6b

δH

δH

δH

δH

δH

3.94, d (4.0) 2.24, m 1.83, m 5.25, dd (7.9, 2.4) 2.18, m 1.58, m 1.81, m 1.42, br dd (14.5, 4.9) 2.21, m 3.58, d (3.5) 1.93, dd (13.7, 12.0) 1.55, m 1.15, s 4.56, d (11.5) 3.64, d (11.5) 1.23, s

3.89, d (4.0) 2.11, m 1.91, m 5.29, dd (8.0, 3.0) 2.17, m 1.57, m 1.79, m 1.35, m 2.11, m 3.58, d (3.5) 1.85, m 1.57, m 1.13, s 4.52, d (11.5) 3.62, d (11.5) 1.18, s

3.88, d (4.2) 2.15, m 1.88, m 5.12, dd (8.0, 3.0) 2.15, m 1.61, m 1.80, m 1.35, br dd (14.3, 5.4) 2.10, m 3.57, d (3.8) 1.85, m 1.57, m 1.15, s 4.52, d (11.6) 3.54, d (11.6) 1.18, s

4.00, d (4.1) 2.38, br dd (15.6, 8.0) 1.84, m 5.19, dd (7.9, 2.2) 2.15, m 1.55, td (14.6, 5.4) 1.78, m 1.40, m 2.16, m 3.62, d (3.6) 1.93, m 1.47, dd (13.9, 5.4) 1.10, s 4.55, d (11.7) 3.65, d (11.7) 1.25, s

3.71, d (5.2) 2.55, dd (8.1, 15.4) 2.02, m 5.88, dd (6, 8.1) 1.92, m 1.91, m 2.00−2.08, m 2.00−2.08, m 5.41, d (3.8) 3.75, d (5.2) 3.03, d (4.1) 2.81, d (4.1) 0.85, s 4.21, br d (12.6) 4.20, br d (12.6) 1.71, s

5.99, br s 3.52, br d (13.4) 2.64, ddd (13.3,12.5,6.6) 4.14, td (11.5, 2.7) 4.01, br dd (11.0, 6.2) 1.92, m 1.65, m 2.74, m 2.32, m 6.55, ddd (11.6, 9.2, 6.3) 5.89, d (11.6) 4.26, br s 3.75, d (5.0) 5.36, d (5.0)

6.23, s 3.60, br d (14.0) 2.43, m 4.13, td (12.5, 2.5) 3.92, m 2.02, m 1.88, m 2.71, m 2.38, m 6.57, m 5.87, dd (11.5, 1.0) 4.61, s 4.39, s

6.06, br s 3.86, br d (13.4) 2.30, m 4.14, t (11.0) 3.91, br dd (11.2, 6.1) 1.97, m 1.90, m 2.75, m 1.58, m 6.51, m 5.80, dd (11.6, 2.3) 4.46, br s

6.00, br s 3.57, d (13.3) 2.53, m 4.23, t (11.7, 2.3) 4.06, dd (11.0, 5.9) 2.05, m 1.70, m 2.83, m 1.70, m 6.42, m 5.78, dd (11.6, 2.1) 4.96, br s

3.72, d (5.3) 2.56, dd (15.4, 8.1) 2.04, m 5.86, m 1.95, m 1.85, m 2.13, m 1.94, m 5.65, dd (4.6) 3.80, (d, 5.0) 3.03, d (4.0) 2.80, d (4.0) 0.82, s 4.21, d (12.5) 3.93, d (12.5) 3.97, m 3.97, m 5.86, br s 3.55, d (12.9) 2.76, m 3.93, m 4.03, m 2.30, ddd (13.6,13.6, 7.8) 1.57, td (13.6, 3.0) 2.76, m 1.67, m 6.53, m 5.78, dd (11.6, 1.8) 4.02, s

2.25, s

2.30, s 2.01, s

6.21, s 3.63, br d (12.9) 2.46, dd (12.6, 5.7) 4.10, t (11.0) 3.91, dd (10.8, 5.5) 2.02, m 2.00, m 2.12, td (4.8, 13.9) 1.78, td (3.5, 13.9) 6.61, m 5.83, dd (1.8, 11.6) 4.65, s 3.81, s

2.21, s

a

Recorded at a Bruker DRX500 in CD3OD. bRecorded at a Bruker DRX500 in acetone-d6. cRecorded at a Bruker DPX300 in acetone-d6. Recorded at a Bruker DRX500 in CDCl3.

d

The 3D contour maps for the CoMFA and CoMSIA models of K562 cells are shown in Figure 3 and indicate the relationship between the structural features and bioactivity. Because of its strong activity (IC50 values to K562 and SW1116 cell lines: 10 and 30 nM), the contours of steric, electrostatic, hydrophobic, and H-bond-accepting factors were aligned for epiisororidin E (21). In the steric contour map (Figure 3a), a large green plot is located around the region of C-6′, C-8′, and C-12′ of the macrocyclic ring of 21, suggesting that the activity would be increased if this area is bulky. Compounds with an expanded ring, such as compounds 13, 14, 17, 18, 21, and 22, show higher activities. In the electrostatic contour map (Figure 3b), the blue contour near the C-1′ and C-11′ positions indicates that the electropositivity around this region would enhance activity, and the red plot at the 12,13-epoxide suggests that an electronegative group at the position is desirable for bioactivity. The yellow contours around C-1′ and C-3′ (Figure 3c) indicate that the hydrophobic groups around this position increase the activity, while the gray plot near the C-6′ demonstrates the positive dependence of activity upon the hydrophilicity near this region, as shown with vertisporin (19) > 14′-dehydrovertisporin (6) and 12′-episatratoxin H (20) > 16-hydroxymytoxin B (5). The gray block over the C-5, C-6, and C-4 regions also reveals a positive dependence of activity upon the hydrophilicity; for

3). Furthermore, the structural diversity of this collection of compounds enabled insight into the structure−activity relationships. Briefly, the 12′,13′-epoxide group contributes to the activity, and removing this decreases cytotoxicity. The 12′-Oacetyl group in compound 4 enhanced bioactivity compared to compound 7. The introduction of 16-OH into roridin E (11) and mytoxin B (12) (compared to compound 16 and compound 5, respectively) reduced the activity remarkably. The cytotoxicity was shown to be chirality-dependent by the difference in bioactivity between roridin E (11) and epiisororidin E (21), as well as between myrothecines A (7) and F (3). The CoMFA (comparative molecular field analysis) and CoMSIA (comparative molecular similarity indices analysis) approaches were coapplied for the subsequent 3D QSAR study,14,15 which would shed light on the optimization for drug leads derived from this type of compound. Vertisporin (19),16 12′-episatratoxin H (20),17 epiisororidin E (21),18 and 16-hydroxyverrucarin A (22),7 isolated by us from Myrothecium sp. Z16 (not reported), were also evaluated for their cytotoxicity for the 3D QSAR study. For K562 and SW1116 cell-based data sets, the CoMFA and CoMSIA 3D QSAR models were generated on the basis of the training sets of 17 compounds (Table S2). The CoMFA and CoMSIA models have good predictabilities (Figure 2), and the pIC50’s for both K562 and SW1116 cells are shown in Table S3. C

DOI: 10.1021/acs.jnatprod.8b01034 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Table 2. 13C NMR Data of Compounds 1−6 (δ in ppm) 1a

2b

3c

4d

5b

6b

no.

δC, type

δC, type

δC, type

δC, type

δC, type

δC, type

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1′ 2′ 3′ 4′ 5′ 6′ 7′ 8′ 9′ 10′ 11′ 12′ 13′ 14′ 15′ 16′

84.6, CH 41.8, CH2 82.9, CH 53.2, C 45.8, C 30.6, CH2 32.6, CH2 74.8, C 46.0, CH 71.6, CH 78.7, C 30.0, CH2 11.8, CH3 75.2, CH2 28.4, CH3 168.1, C 119.9, CH 154.8, C 27.9, CH2 66.0, CH2 85.9, C 30.8, CH2 23.8, CH2 149.9, CH 123.1, CH 168.6, C 85.3, CH 80.3, CH 105.3, CH

81.6, CH 40.8, CH2 79.3, CH 52.0, C 44.7, C 28.8, CH2 31.9, CH2 72.9, C 45.0, CH 70.0, CH 77.4, C 30.3, CH2 10.9, CH3 74.2, CH2 28.0, CH3 166.0, C 121.5, CH 149.0, C 26.2, CH2 64.4, CH2 82.4, C 29.0, CH2 22.5, CH2 148.4, CH 122.3, CH 166.9, C 81.1, CH 76.8, CH 175.6, C

81.6, CH 41.4, CH2 79.4, CH 51.9, C 44.5, C 29.0, CH2 31.9, CH2 72.9, C 45.1, CH 70.1, CH 77.5, C 30.6, CH2 11.0, CH3 73.9, CH2 27.9, CH3 167.2, C 114.2, CH 156.8, C 31.4, CH2 63.9, CH2 87.8, C 25.0, CH2 21.2, CH2 148.5, CH 122.1, CH 167.1, C 74.0, CH 209.6, C 25.3, CH3

80.8, CH 40.4, CH2 79.2, CH 51.3, C 44.0, C 28.2, CH2 31.3, CH2 73.6, C 44.2, CH 69.0, CH 78.1, C 28.2, CH2 10.3, CH3 73.0, CH2 28.1, CH3 165.3, C 120.1, CH 149.6, C 25.9, CH2 63.2, CH2 85.7, C 29.5, CH2 21.1, CH2 147.9, CH 120.9, CH 165.8, C 77.3, CH 211.6, C 28.0, CH3 169.3, C 20.7, CH3

79.6, CH 35.6, CH2 74.6, CH 50.4, C 44.2, C 20.9, CH2 23.6, CH2 143.5, C 119.2, CH 67.7, CH 66.0, C 47.6, CH2 8.30, CH3 64.8, CH2 65.6, CH2 167.1, C 117.1, CH 156.6, C 26.4, CH2 63.9, CH2 88.7, C 28.5, CH2 22.7, CH2 150.4, CH 121.6, CH 166.6, C 78.1, CH 212.3, C 29.0, CH3

79.6, CH 35.5, CH2 74.6, CH 50.4, C 43.8, C 21.1, CH2 28.1, CH2 139.5, C 120.5, CH 67.9, CH 66.0, C 47.6, CH2 8.30, CH3 65.0, CH2 23.2, CH3 166.5, C 120.4, CH 150.0, C 26.1, CH2 65.6, CH2 73.3, C 22.5, CH2 22.6, CH2 150.7, CH 121.7, CH 166.7, C 84.6, CH 73.2, CH 175.9, C

a

Recorded at a Bruker DRX500 in CD3OD. bRecorded at a Bruker DRX500 in acetone-d6. cRecorded at a Bruker DPX300 in acetone-d6. Recorded at a Bruker DRX500 in CDCl3.

d

cells (Figure 4). Figure 4a shows the steric contour maps with the yellow contour near C-1′ and C-3′, suggesting that reduction of bulky groups at the position would increase the activity, and the green contour near C-6′, C-2, and C-5 indicated that bulky groups at this position would increase the activity. The electrostatic contour maps are displayed in Figure 4b, in which blue contour near the C-13′, C-3′, and C-9′ positions suggests that the activity would be enhanced by the electropositivity. The theoretical prediction regarding the steric, electrostatic, hydrophobic, and H-bond-accepting factors agrees well with the experimental data, underpinning its guiding value in the drug lead optimization starting from this collection of macrolides.

■

Figure 1. X-ray crystal structure of myrothecine D (1, five H2O molecules omitted for clarity).

EXPERIMENTAL SECTION

General Experimental Procedures. Melting points were measured on an XT-4 apparatus and are uncorrected. Optical rotation was determined in MeOH on a WXG-4 disc polarimeter, and IR spectra in KBr disks on a Nexus 870 FT-IR spectrometer. The UV spectra were recorded on a Hitachi U-3000 spectrophotometer. NMR spectra were acquired on Bruker DRX-500 and DPX-300 NMR spectrometers using solvent signals as internal standards. HR-ESIMS were taken on a Mariner Mass 5304 instrument. CD spectra were recorded on a Jasco J810 circular dichroism spectrometer. The ELISA plate reader was from Sunrise, USA. HPLC were performed with an Apollo C18 (250 × 4.6 mm, 5 μm) column, a Hitachi pump L-7100, and a UV detector L-7400.

example, the cytotoxicity of 16-hydroxyverrucarin A (22) is higher than that of roridin A (13), 12′-hydroxyroridin E (10), and 16-hydroxyroridin E (16). In the CoMSIA acceptor contour map (Figure 3d), the magenta contours located at the ester carbonyl C-1′ and C11′ indicate the hydrogen bond acceptor would increase the activity, while the red at the C-16 position illustrates that the activity would be weakened by a H-bondaccepting substituent. The steric and electrostatic contour maps were generated for vertisporin (19) with the model of SW1116 D

DOI: 10.1021/acs.jnatprod.8b01034 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

MeOH−H2O (50:50, 1.2 mL/min) to give 3 (8 mg, tR = 15.5 min). The HPLC separation of Fr-3-3-3 using MeOH−H2O (60:40, 1.0 mL/min) afforded 4 (6 mg, tR = 14.2 min). The dried extract (19.8 g) derived from IFB-E008 was first fractionated as described elsewhere.1 Fr-5 was purified by silica gel column chromatography with a CHCl3−MeOH gradient (100:0 → 100:16) followed by gel filtration over Sephadex LH-20 with CHCl3−MeOH (1:1) to yield 1 (4.7 mg). The dried extract derived from IFB-E009 was separated as outlined earlier.1 The obtained Fr-2 was chromatographed over a silica gel column with CHCl3−MeOH (100:1 → 100:8), and the afforded subfraction Fr-2-3 was repeatedly purified over a Sephadex LH-20 column using CHCl3− MeOH (1:1) to provided 6 (6 mg). Myrothecine D (1): colorless needles; mp 195−197 °C; [α]25 D −35.5 (c 0.28 CH3OH); UV (MeOH) λmax (log ε) 213 (4.21) nm; CD (CH3OH) λmax (Δε) 234 (−4.30), 252 (+0.36) nm; IR (KBr) νmax 3408, 2962, 2926, 1714, 1657, 1635, 1413, 1397, 1253, 1191, 1061, 1024, 996 cm−1; 1H and 13C NMR, see Tables 1 and 2; HR-ESIMS m/z 585.2305 [M + Na]+ (calcd for C29H38O11Na, 585.2306). Myrothecine E (2): colorless crystals; mp 258−260 °C; [α]25 D +47.1 (c 0.40 CH3OH); UV (MeOH) λmax (log ε) 219 (4.62) nm; CD (CH3OH) λmax (Δε) 234 (−3.06), 249 (+0.57), 270 (−0.20) nm; IR (KBr) νmax 3393, 2961, 2928, 1713, 1659, 1636, 1414, 1249, 1180, 1156, 1056, 1024, 774 cm−1; 1H and 13C NMR data, see Tables 1 and 2; HR-ESIMS m/z 583.2151 [M + Na]+ (calcd for C29H36O11Na, 583.2150). Myrothecine F (3): white powder; mp 160−163 °C; [α]25 D −15.0 (c 0.53 CH3OH); UV (MeOH) λmax (log ε) 219 (4.13) nm; CD (CH3OH) λmax (Δε) 235 (−5.60), 255 (−0.06), 296 (−4.02) nm; IR (KBr) νmax 3445, 2958, 2919, 1713, 1692, 1650, 1414, 1385, 1250, 1182, 1154, 1100, 1057, 1022, 819, 744 cm−1; 1H and 13C NMR data, see Tables 1 and 2; HR-ESIMS m/z 569.2353 [M + Na]+ (calcd for C29H38O10Na, 569.2357). Myrothecine G (4): colorless crystals; mp 250−252 °C; [α]25 D +9.1 (c 0.33 CH3OH); UV (MeOH) λmax (log ε) 220 (4.20) nm; CD (CH3OH) λmax (Δε) 230 (−4.97), 248 (+1.20), 293 (−0.41) nm; IR (KBr) νmax 3500, 2960, 2922, 2852, 1748, 1717, 1661, 1413, 1227, 1181, 1154, 1061, 1923, 753 cm−1; 1H and 13C NMR data, see Tables 1 and 2; HR-ESIMS m/z 611.2468 [M + Na]+ (calcd for C31H40O11Na, 611.2463). 16-Hydroxymytoxin B (5): colorless crystals; mp 195−198 °C; [α]25 D +30.1 (c 1.26 CH3OH); UV (MeOH) λmax (log ε) 220 (4.18) nm; CD (CH3OH) λmax (Δε) 232 (−1.50), 248 (+3.75), 289 (−1.45) nm; IR (KBr) νmax 3408, 2962, 1714, 1640, 1413, 1384, 1250, 1231, 1179, 1083, 1056, 1026, 1006, 971, 818, 772 cm−1; 1H and 13C NMR data, see Tables 1 and 2; HR-ESIMS m/z 567.2197 [M + Na]+ (calcd for C29H36O10Na, 567.2201). 14′-Dehydrovertisporin (6): colorless crystals; mp 197−199 °C; [α]25 D −25 (c 0.20 CH3OH); UV (MeOH) λmax (log ε) 215 (4.40) nm; CD (CH3OH) λmax (Δε) 232 (−2.82), 248 (+1.66) nm; IR (KBr) νmax 3436, 2966, 1794, 1713,1636, 1437, 1411, 1248, 1185, 1092, 993, 967, 818 cm−1; 1H and 13C NMR data, see Tables 1 and 2; HR-ESIMS m/z 565.2047 [M + Na]+ (calcd for C29H34O10Na, 565.2044). Single-Crystal X-ray Diffraction. Crystal structure determination of 1 was carried out on a Bruker SMART APEX CCD diffractometer equipped with graphite-monochromated Mo Kα (λ = 0.710 73) radiation with Lorentz polarization and absorption corrections for a crystal of 1 (0.35 mm × 0.20 mm × 0.21 mm). The intensities were collected at 293 K using the ω-scan mode with variable scan speed. A total of 3440 reflections were collected in the range of θ = 1.17−25.97°, of which 1679 were independent, which were used in the structure solution and refinements. The structure was solved by direct methods and refined on F2 by full-matrix least-squares methods using SHELX97. All the non-hydrogen atoms were refined anisotropically. All the hydrogen atoms were placed in calculated positions and were assigned fixed isotropic thermal parameters at 1.2 times the equivalent isotropic U of the atoms to which they are attached and allowed to ride on their respective parent atoms. The contributions of these hydrogen atoms were included in the structure-factor calculations. The refinement gave the final R1 = 0.1021 with w = [σ2(Fo)2 + (0.1(max(0,Fo2) + 2Fc2)/

Table 3. Cytotoxicity of Compounds 1−22 Expressed in IC50 Values (μM) compound

K562

SW1116

myrothecine D (1) myrothecine E (2) myrothecine F (3) myrothecine G (4) 16-hydroxymytoxin B (5) 14′-dehydrovertisporin (6) myrothecine A (7) myrothecine B (8) myrothecine C (9) 12′-hydroxyroridin E (10) roridin E (11) mytoxin B (12) roridin A (13) verrucarin A (14) 14′-hydroxymytoxin B (15) 16-hydroxyroridin E (16) 16-hydroxyverrucarin J (17) trichoverritone (18) vertisporin (19) 12′-episatratoxin H (20) epiisororidin E (21) 16-hydroxyverrucarin A (22) 5-fluorouracil (control)

8.20 15.98 0.97 1.53 2.87 0.056 31.59 26.69 15.95 3.15 0.54 0.0013 0.0071 0.0030 0.92 7.13 0.0064 0.0012 0.0011 0.51 0.0010 0.0017 32.92

0.57 11.61 10.62 4.25 0.18 0.20 22.71 23.13 21.07 7.92 0.039 0.0015 0.21 0.0018 0.55 9.06 0.16 0.0031 0.0015 2.27 0.033 0.012 76.92

Silica gel (200−300 mesh) for column chromatography and silica GF254 for TLC were produced by Qingdao Marine Chemical Company, China. Sephadex LH-20 was purchased from Pharmacia Biotech, Sweden. ODS silica gel was from Nacalai Tesque, Kyoto, Japan. All chemicals used in the study were of analytical grade. Endophytic Strains. The M. roridum IFB-E008 and IFB-E009 were separated from T. jasminoides (Apocynaceae), and IFB-E012 from A. annua (Asteraceae). The three fungal strains were cultured on PDA medium on Petri dishes for morphologic observation. The spore masses (colony) look viscous and green when young and turn hard and black while aging. Conspicuous sporodochia were formed on densely compacted branching conidiophores, with the ultimate branches being phialides. Phialides were cylindrical with conically tapering tips and undifferentiated collarettes. Conidia were olive-brown, singlecelled, and cylindrical with both ends rounded ((5.2−7.3) × 2.0 μm). The identification of these M. roridum strains was reinforced by their 18S rDNA sequences, which showed a 99% similarity to those accessible at the BLASTN of M. roridum. The sequences of M. roridum IFB-E008, IFB-E009, and IFB-E012 have been deposited as EF211124, EF211125, and DQ102373 in GenBank, respectively. The live cultures of the three M. roridum strains were kept at the Institute of Functional Biomolecules, Nanjing University (China). Isolation of New Macrolides (1−6). The endophytic strains were cultured as detailed previously,1 and the obtained biomasses were extracted exhaustively with EtOAc or MeOH, respectively. The dried extract derived from IFB-E012 was fractionated into six parts (Fr-1−Fr6), and Fr-3 was subjected to further column chromatography over silica gel with a CHCl3−MeOH gradient (100:0 → 100:16). The macrolide-containing subfraction traced by 1H NMR was further separated through gel filtration over Sephadex LH-20 with CHCl3− MeOH (1:1) followed by ODS column chromatography eluting with H2O−MeOH gradient (100:0 → 80:20) and purified by semipreparative HPLC using MeOH−H2O (50:50, 1 mL/min) to yield 2 (6 mg, tR = 6 min). Gel filtration of the subfraction Fr-3-3 over Sephadex LH-20 with CHCl3−MeOH (1:1) followed by silica gel column chromatography with CHCl3−MeOH gradient (100:0 → 100:16) gave Fr-3-3-1, Fr-3-3-2, and Fr-3-3-3. Fr-3-3-1 was further purified by HPLC using MeOH−H2O (47:53, 1.2 mL/min) to afford 5 (10 mg, tR = 32 min). Fr-3-3-2 was further purified by HPLC using E

DOI: 10.1021/acs.jnatprod.8b01034 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Figure 2. Graphs of the experimental versus predicted pIC50 values of compounds in training and test sets for CoMFA and CoMSIA models. (a) and (b) represent activities in K562 cells, (c) and (d) denote activities in SW1116 cells. Triangles and circles represent compounds in the training and test sets, respectively. 3)2]−1.12 Crystallographic and experimental data for myrothecine D (1) are listed in Table S1. Cytotoxicity Assay. The in vitro cytotoxicity against K562 and SW1116 cell lines was evaluated as described earlier.1 3D QSAR. 3D QSAR models were developed using the data set of 22 trichothecene macrolides. The data set was divided into a training set of 17 and a test set of five compounds, respectively. The activity (i.e., IC50 value) of these macrocylic compounds against K562 and SW1116 cells were converted to the corresponding pIC50’s (= −log IC50’s) (Table S3). All the compounds’ partial charges were calculated using the Gasteiger−Huckel method, and their optimal geometry was done using the Tripos force field with a distance-dependent dielectric function and an energy convergence criterion of 0.05 kcal/mol Å using 100 iterations, which were done using the SYBYL6.9 program (Tripos Inc.). It is a crucial step to perform the alignment of molecules in the 3D QSAR study. In order to obtain a consistent alignment, the lowest energy conformation of the most active compound in the series was used as the template; therefore, compounds epiisororidin E (21) and vertisporin(19) were selected as the templates of SW1116 and K562 cells, respectively. The aligned compound results are similar and are shown inFigure S2.

CoMFA and CoMSIA are considered be the most reliable methods for the 3D QSAR study. The Tripos Sybyl 6.9 program was used to perform CoMFA and CoMSIA modelings. Since the experimental K562 and SW1116 activities varied significantly, different test and training sets were used to develop the 3D QSAR models. The CoMFA models were developed with steric and electrostatic fields, while the CoMSIA models also explored the impacts of more fields such as hydrophobic, hydrogen bond donor, and hydrogen bond acceptor in addition to steric and electrostatic fields. The partial least-squares analysis method was used to derive the 3D QSAR models, describing the discrimination between independent variables (the descriptors of CoMFA and CoMSIA) and dependent variables (the pIC50 values). The leave-one-out cross-validation was carried out to obtain the optimal number of components and the highest correlation coefficient, which was further used to obtain the final 3D QSAR model. The robustness of the developed models was evaluated with various parameters such as non-cross-validation correlation coefficient, standard error of estimate, and Fischer ratio value, and the predictive ability was further tested by compounds in the test set. F

DOI: 10.1021/acs.jnatprod.8b01034 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

■

trichothecene macrolide data set, and the proposed biosynthetic pathway (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +86-514-87992233. ORCID

Li Shen: 0000-0003-1947-5236 Ren-Xiang Tan: 0000-0001-6532-6261 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The work was cofinanced by the National Natural Science Foundation of China (21372191, 81121062, 21132004, and 21072092).

■

(1) Shen, L.; Jiao, R. H.; Ye, Y. H.; Wang, X. T.; Xu, C.; Song, Y. C.; Zhu, H. L.; Tan, R. X. Chem. - Eur. J. 2006, 12, 5596−5602. (2) Liu, J. Y.; Huang, L. L.; Ye, Y. H.; Zou, W. X.; Guo, Z. J.; Tan, R. X. J. Appl. Microbiol. 2006, 100, 195−202. (3) Amagata, T.; Rath, C.; Rigot, J. F.; Tarlov, N.; Tenney, K.; Valeriote, F. A.; Crews, P. J. Med. Chem. 2003, 46, 4342−4350. (4) Matsumoto, M.; Minato, H.; Tori, K.; Ueyama, M. Tetrahedron Lett. 1977, 47, 4093−4096. (5) Jarvis, B. B. Bull. Soc. Chim. Belg. 1986, 95, 681−697. (6) Jarvis, B. B.; Midiwo, J. O.; Mazzola, E. P. J. Med. Chem. 1984, 27, 239−244. (7) Jarvis, B. B.; Stahly, G. P.; Pavanasasivam, G.; Mazzola, E. P. J. Med. Chem. 1980, 23, 1054−1058. (8) Xu, J.; Takasaki, A.; Kobayashi, H.; Oda, T.; Yamada, J.; Mangindaan, R. E. P.; Ukai, K.; Nagai, H.; Namikoshi, M. J. Antibiot. 2006, 59, 451−455. (9) Breitenstein, W.; Tamm, C. Helv. Chim. Acta 1975, 58, 1172− 1180. (10) Alvi, K. A.; Rabenstein, J.; Woodard, J.; Baker, D. D.; Bergthold, J. D.; Lynch, J.; Lieu, K. L.; Braude, I. A. J. Nat. Prod. 2002, 65, 742−744. (11) Jarvis, B. B.; Vrudhula, V. M.; Pavasasivam, G. Tetrahedron Lett. 1983, 24, 3539−3542. (12) Crystallographic data in CIF format have been deposited in the Cambridge Crystallographic Data Centre as CCDC-951565 [available free of charge at http://www.ccdc.cam.ac.uk/deposit or from the CCDC, 12 Union Road, Cambridge CB21EZ, UK, fax: (+44) 1223336-033; or e-mail: [email protected]]. (13) Kunishima, M.; Ujigawa, T.; Nagaoka, Y.; Kawachi, C.; Hioki, K.; Shiro, M. Chem. - Eur. J. 2012, 18, 15856−15867. (14) Gangwal, R. P.; Bhadauriya, A.; Damre, M. V.; Dhoke, G. V.; Sangamwar, A. T. Curr. Top. Med. Chem. 2013, 13, 1015−1035 and references therein. . (15) Ai, C. Z.; Li, Y.; Wang, Y. H.; Chen, Y. D.; Yang, L. Bioorg. Med. Chem. Lett. 2009, 19, 803−806. (16) Minato, H.; Katayama, T.; Tori, K. Tetrahedron Lett. 1975, 30, 2579−2582. (17) Piao, M. Z.; Shen, L.; Wang, F. W. J. Asian Nat. Prod. Res. 2013, 15, 1284−1289. (18) Jarvis, B. B.; Wang, S. J. J. Nat. Prod. 1999, 62, 1284−1289.

Figure 3. Contour maps derived from 3D QSAR model of K562 cells. (a) The green and yellow represent the favorable and unfavorable steric fields, respectively. (b) The blue and red denote electropositive and electronegative fields, respectively. (c) The yellow and gray indicate the favorable and unfavorable hydrophobic fields, respectively. (d) The magenta and red describe the favorable and unfavorable hydrogen bond acceptor fields, respectively. Epiisororidin E (21) was overlaid in each plot.

Figure 4. Contour maps generated from the 3D QSAR model of SW1116. (a) The favorable and unfavorable steric fields are denoted by green and yellow plots, respectively. (b) The blue and red represent electropositive and electronegative fields, respectively. Vertisporin (19) was overlaid in each map.

■

REFERENCES

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jnatprod.8b01034. CD and NMR spectra of 1−6, structures of compounds 13−22, crystallographic and experimental data for 1, Xray molecular structure and X-ray packing diagram of 1, observed and predicted pIC50’s for 1−22, statistical results of CoMFA and CoMSIA models, alignment of the G

DOI: 10.1021/acs.jnatprod.8b01034 J. Nat. Prod. XXXX, XXX, XXX−XXX