Le Yia,
Cheng-Bin Cui*a,
Chang-Wei Lia,
Ji-Xing Pengb and
Qian-Qun Gub
aState Key Laboratory of Toxicology and Medical Countermeasures, Beijing Institute of Pharmacology and Toxicology, Beijing 100850, China. E-mail: cuicb@126.com; cuicb@sohu.com; Fax: +86-10-68211656; Tel: +86-10-68211656
bKey Laboratory of Marine Drugs, Chinese Ministry of Education, School of Medicine and Pharmacy, Ocean University of China, Qingdao 266003, China
First published on 29th April 2016
Chromosulfine (1), a novel cyclopentachromone sulfide, was isolated from a neomycin-resistant mutant of the marine-derived fungus, Penicillium purpurogenum G59. Its structure including stereochemistry was determined using spectroscopic methods, in particular NMR and electronic CD (ECD) analysis and Mosher's method. HPLC-UV/HPLC-MS analyses demonstrated that 1 was produced only in the mutant by a biosynthetic pathway that is silent in the parent strain and was activated by the introduction of neomycin resistance.
Because it is known that major biosynthetic gene clusters for secondary metabolite production in bacteria and fungi are generally silenced under standard culture conditions, various strategies have been developed to activate these gene clusters over the past decade,7 and the activation of silent biosynthetic pathways for secondary metabolites has become a promising route to discover new compounds. Due to the culture-based simple procedures involved, the one-strain-many-compounds,8 chemical epigenetics,9 and cocultivation10 strategies have been widely used for this purpose, particularly by microbial chemists. A recently reported mutagenesis strategy using diethyl sulfate also provides a simple and practical method for microbial chemists to discover new compounds derived from silent fungal pathways.2b,11 The ribosome engineering strategy for the production of new compounds, which has been well-studied in bacteria12 and can activate silent bacterial pathways by introducing drug resistance,13 has been extended recently to fungi by our group particularly with the development of new methodologies specific to fungi.14
We previously reported that drug resistance in fungi can be introduced with antibiotics using DMSO or ultrasound,14e,14f resulting in the production of diverse secondary metabolites in the mutants by activating pathways that are silent in the parental strain.14 We have also shown that a neomycin-resistant mutant (4-30) selected by treating marine-derived Penicillium purpurogenum G59 spores with 6.7 mg mL−1 neomycin in 67% DMSO at 4 °C for 7 days produced certain polyketides and steroids in liquid culture by activating pathways that are silent in the G59 strain.14f In a continuation of these studies, we report herein a novel and the first CPC carrying a sulfide chain, chromosulfine (1, Fig. 1), produced in solid culture by the 4-30 mutant by activating a pathway that is silent in the G59 strain by generating drug resistance.
In this study, the mutant 4-30 and parental G59 strains were concurrently fermented using rice as a solid-substrate fermentation medium at 28 °C for 80 days under the same conditions instead of liquid-medium fermentation as used previously,14f prior to obtaining ethyl acetate (EtOAc) extracts from the cultures. Similar to bioassays using liquid-medium fermentation,14 extracts from the mutant inhibited K562 cells with an inhibition rate (IR%) of 51.8% at 100 μg mL−1, whereas the parent extract exhibited no inhibition with an IR% of 5.9% at 100 μg mL−1. A chromatographic separation of the mutant extract, which traced the newly produced metabolites in the mutant by direct comparison with the parent-derived extract using both bioassays and HPLC analysis of typical UV absorption peaks, resulted in the isolation of 1.
Chromosulfine (1) was obtained as a brown amorphous powder with MeOH, [α]25D −14.9 (c 0.5, MeOH), and the molecular formula, C19H20O9S, was assigned by HRESIMS (m/z 447.0728 [M + Na]+, calculated for C19H20NaO9S 447.0726) combined with the 1H and 13C NMR data (Table 1). The typical UV absorption peaks in MeOH at 228 (logε 4.19), 238 (4.19), 258 (4.04) and 323 nm (3.56) indicated a chromone-chromophore in the molecule.2b The IR absorption indicated hydroxyl (3274 cm−1), ester (1737 cm−1), and conjugated carbonyl (1656 cm−1) groups, but the presence of a thiol group was excluded by the absence of IR absorption from S–H bonds in the 2600–2540 cm−1 region.15 The 1H NMR spectrum of 1 in DMSO-d6 exhibited signals from one tert-methyl, two methoxy, two isolated aromatic, and three hydroxyl protons together with several methine and methylene proton signals (Table 1). The chemical shifts of the tert-methyl (δH 2.38) and one of the OH protons (δH 12.34) as well as the signal patterns of the two isolated aromatic protons (δH 6.94/6.96, both br s) revealed the chromone unit with CH3 and OH at the para- and peri-positions of the carbonyl group.1,2,5 The 13C NMR data were analyzed with DEPT and also supported the chromone unit with signals ascribable to one methyl, two sp2 methines, three each of quaternary and oxygen-bearing quaternary sp2 carbons, and one conjugated carbonyl carbon in the unit (Table 1). The signals of the carbonyl at δC 178.2 and the two sp2 carbons at δC 171.8 and 121.6 were consistent with the data for the α,β-unsaturated carbonyl moiety in the CPC system.1,2b,4 The 13C NMR data further indicated the presence of two ester carbonyls, one oxygen-bearing quaternary sp3 carbon, and two each of the methoxy, methylene and methine groups (Table 1). Further detailed analysis of the 1H and 13C NMR data by 2D 1H–1H COSY, HMBC, HMQC, ROESY and 1D GOESY difference NOE spectroscopy established the planar structure of 1 according to the key NMR data provided in Fig. 2.
Position | δHa (J in Hz) | δCa | HMBC, H → C |
---|---|---|---|
a The δH/δC values were recorded in DMSO-d6 using the DMSO-d6 signals (δC 39.52/δH 2.50) as references.b The signal disappeared after adding drops of D2O. | |||
1 | — | 172.2 s | — |
2 | — | 79.4 s | — |
3 | 3.98 dd (9.0, 8.1) | 51.0 d | C-1,2,4,13,16 |
4 | Hα 3.06 dd (17.4, 9.0) | 37.5 t | C-3,5,6,13 |
Hβ 3.46 dd (17.4, 8.1) | C-2,3,5,13 | ||
5 | — | 121.6 s | — |
6 | — | 178.2 s | — |
7 | — | 108.2 s | — |
8 | — | 159.9 s | — |
9 | 6.94 br s | 112.4 d | C-6,7,8,11,14 |
10 | — | 147.3 s | — |
11 | 6.96 br s | 108.0 d | C-6,7,9,12,14 |
12 | — | 156.5 s | — |
13 | — | 171.8 s | — |
14 | 2.38 3H, br s | 21.7 q | C-9,10,11 |
15 | 3.71 3H, s | 52.5 q | C-1 |
16 | Ha 2.86 dd (13.7, 5.1) | 34.7 t | C-3,17,18 |
Hb 2.78 dd (13.7, 6.9) | C-3,17,18 | ||
17 | 4.22 dd (6.9, 5.1) | 70.7 d | C-16,18 |
18 | — | 172.7 s | — |
19 | 3.66 3H, s | 51.6 q | C-18 |
2-OH | 5.83b br s | — | |
8-OH | 12.34b br s | — | |
17-OH | 5.95b br s | — | C-7,8,9 |
Because no NOE in ROESY and GOESY was informative, the relative configuration at C-2 and C-3 in 1 was determined by the 3JC,H values of C-1 and H-3. We measured the 1H-coupled 13C NMR spectrum of 1 with selective 1H decoupling of H3-15 protons in DMSO-d6 after adding drops of D2O to fully exchange all of the OH signals. Under this condition, the long-range C–H coupling of C-1 with HO-2 and H3-15 was removed by exchanging and decoupling. Thus, the C-1 signal in the spectrum split into a doublet by coupling with only H-3, as shown in Fig. 3, generating the coupling constant of 2.4 Hz for C-1 and H-3. According to the Karplus relation,16 two possible dihedral angles between the C-1–C-2 and C-3–H-3 bonds, approximately 56° and 115°, respectively, were derived from the 2.4 Hz coupling constant. The former angle of 56° was excluded because this formation is impossible on a restricted cyclopentene ring with an envelope conformation.17 However, the orientation of C-1 and H-3 on the opposite side of the cyclopentene ring gives rise to the 115° dihedral angle. Thus, the relative configuration at C-2 and C-3 in 1 was determined.
![]() | ||
Fig. 3 Enlarged region of the 1H-coupled 150 MHz 13C NMR spectrum of 1 with selective decoupling of H3-15 in DMSO-d6 with drops of D2O. |
The absolute configuration of C-2 and C-3 in 1 was determined by TDDFT electronic CD (ECD) calculations18 in conjunction with the measured CD data (Fig. 4). In the ECD calculations, an attempt at searching conformations, which were run on the full structure of 1 using the ‘systematic’ procedure implemented in Spartan'14 using a Merck molecular force field (MMFF),19 failed to identify the MMFF minima. Therefore, additional calculations were performed on simplified model structures by removing an entire chiral centre in the side chain,18a 1s and ent-1s shown in Fig. 4, because no significant effect of side chain chirality on the CD of the chiral CPC ring system in 1 was expected in view of the far location of chiral carbon in the side chain from the CPC ring system.18a All of the MMFF minima from the conformational searches19 for 1s and ent-1s were re-optimized according to the DFT calculations at the B3LYP/6-31+G(d) level using the Gaussian 09 software package.20 TDDFT ECD calculations were performed on a set of two lowest-energy conformations (>5% population) for 1s and ent-1s each (ESI, Fig. S1 and S2†) with 30 excited states using a polarizable continuum model (PCM) in acetonitrile. The CD spectra were generated with the SpecDis21 program by applying a Gaussian band shape with a 0.35 eV width from dipole-length rotational strengths. The spectra of the conformers were combined according to the Boltzmann weighting for each conformer to obtain the ECD spectra of 1s and ent-1s. As shown in Fig. 4, the calculated ECD spectrum of 1s properly reproduced the measured CD data of 1 in acetonitrile except for an inverted peak at 208 nm in the calculated spectrum and an unmatched additional negative Cotton effect at 317 nm in the measured spectrum. These two peaks could be ascribed to the π → π* and n → π* transitions of the ester carbonyl that directly linked with the chiral C-17 carbon with an OH in the side chain of 1. Thus, the absolute configuration at C-2 and C-3 in 1 was assigned as 2S,3R, which is the same as 1s.
![]() | ||
Fig. 4 Measured and calculated ECD spectra of 1 and its simplified model structure compounds, 1s and ent-1s, in acetonitrile. |
The absolute configuration of C-17 in 1 was then determined by the modified Mosher's method.22 The treatment of 1 with (S)- and (R)-MTPA Cl in pyridine-d5 in NMR tubes, which allowed monitoring of the esterification reaction with 1H NMR, generated 1H NMR data for the (S)- and (R)-MTPA monoesters 1a and 1b in Fig. 5. The 17R configuration of 1 was assigned using the Δδ (δS − δR) values calculated for 1a and 1b (Fig. 5).
To determine whether 1 was also produced in the parental G59 strain, the EtOAc extracts from the mutant 4-30 and the parental G59 strain were analyzed with HPLC-photodiode array detector-UV scanning and HPLC-ESI-MS using compound 1 as a reference standard. In the HPLC-UV and HPLC-MS analysis, compound 1 was detected only in the mutant extract and not the parent extract as determined by retention times and both the UV and MS spectra (ESI, Fig. S3 and S4†). This result provided evidence that the production of 1 in the mutant 4-30 strain was caused by the activation of biosynthetic pathways in the mutant that are silent in the parental G59 strain by introducing neomycin resistance. A plausible biosynthetic pathway for 1 that is likely activated in the mutant is proposed in Scheme 1.
The biosynthetic precursor of 1 is proposed to be 2, and the methyl carboxylate of 2 has been isolated from fungal metabolites.23 The intermediate I-1, formed from 2 as proposed for coniochaetones,1d would undergo intramolecular cyclization followed by dehydration and decarboxylation coupled with the following rearrangement of the double bond in I-2 to generate I-3. A conjugated nucleophilic addition of the cysteine metabolite 3-mercaptolactate to the double bond of I-3 (ref. 24) followed by further additional modifications could produce 1 (Scheme 1). Both (R)- and (S)-3-mercaptolactates are created from cysteine metabolism,24a and both of their sulfide derivatives have also been reported from fungal metabolites.24
Compound 1 inhibited the human cancer cell lines K562, HL-60, BGC-823, HeLa and MCF-7 with the IC50 values listed in Table 2, which was determined with the MTT method after treatment of the cells with 1 for 24 h at 37 °C. 5-Fluorouracil, used as a positive control in the MTT test, inhibited these cells with IR% values of 42.2% (K562), 42.4% (HL-60), 46.8% (BGC-823), 42.9% (HeLa), and 45.3% (MCF-7) at 100 μg mL−1. When examined using an inverted microscope, some of the cells treated with 100 μg mL−1 1 for 24 h at 37 °C exhibited apoptotic changes in cellular morphology, including cell desquamation and rounding, typical of apoptosis in adherent cells (BGC-823, HeLa and MCF-7), and apoptotic bodies, typical of apoptosis in suspension cells (K562 and HL-60).
Cell lines | K562 | HL-60 | BGC-823 | HeLa | MCF-7 |
---|---|---|---|---|---|
a Cells treated with the sample for 24 h at 37 °C were assayed using the MTT method. | |||||
IC50 (μM) | 60.8 | 16.7 | 73.8 | 75.4 | 59.2 |
Footnote |
† Electronic supplementary information (ESI) available: Experimental procedures, 1H NMR data for 1a/1b, DFT-optimized structures for the low-energy conformers of 1s and ent-1s, figures for the HPLC-UV and HPLC-MS analyses, ESIMS, HRESIMS, UV, IR, and 1D and 2D NMR spectra for 1. See DOI: 10.1039/c6ra06250d |
This journal is © The Royal Society of Chemistry 2016 |