Biosynthetic Studies on Acetosellin and Structure

Original Paper

Biosynthetic Studies on Acetosellin and Structure Elucidation of a New Acetosellin Derivative*

Authors

Correspondence

Peter Hufendiek 1, Simon Stephan Martin Stölben 1, Stefan Kehraus 1,

Prof. Dr. Gabriele M. König

Nicole Merten 1, Henrik Harms 1, Max Crüsemann 1, Idris Arslan 4,

Institute for Pharmaceutical Biology, University of Bonn

Michael Gütschow 2, Tanja Schneider 3, Gabriele Maria König 1

Nussallee 6, 53115 Bonn, Germany Phone: + 49 22 8 73 37 47, Fax: + 49 22 8 73 32 50

1

Institute for Pharmaceutical Biology, University of Bonn, Bonn,

2 3

Germany Pharmaceutical Chemistry I, University of Bonn, Bonn, Germany Institute for Medical Microbiology, Immunology and Parasitology (IMMIP), Pharmaceutical Microbiology Unit, University of Bonn,

4

Bonn, Germany Biomedical Engineering, Pamukkale University, Denizli, Turkey

[email protected] Supporting information available online at http://www.thieme-connect.de/products

A B S T R AC T Natural products from fungi, especially Ascomycota, play a major role in therapy and drug discovery. Fungal strains originating from marine

Key words Acetosellin, Epicoccum nigrum, Didymellaceae, azaphilones, polyketides, isotope labeling

habitats offer a new avenue for finding unusual molecular skeletons. Here, the marine-derived fungus Epicoccum nigrum (strain 749) was found to produce the azaphilonoid compounds acetosellin and 5′,6′-

received

October 7, 2016

dihydroxyacetosellin. The latter is a new natural product. The biosyn-

revised

December 13, 2016

thesis of these polyketide-type compounds is intriguing, since two

accepted

December 16, 2016

polyketide chains are assembled to the final product. Here we performed

Bibliography

13

C labeling studies on solid cultures to prove this hypothesis

for acetosellin biosynthesis.

DOI http://dx.doi.org/10.1055/s-0042-124493 Published online | Planta Med © Georg Thieme Verlag KG Stuttgart · New York | ISSN 0032‑0943

compound 1 and, consequently, also 2 are outstanding, since they seem to be composed of two polyketide chains. Using 13C-labeled precursors, this study set out to prove this hypothesis.

Introduction Acetosellin and its dihydroxy-derivative (1, 2) belong to a group of fungal polyketides called azaphilonoid compounds [1, 2]. For these, a highly oxidized bicyclic pyran quinone core structure is characteristic, with the oxygen being substituted by nitrogen for some members of this class (▶ Fig. 1). A remarkable feature of the acetosellins, when compared to the azaphilones in general, is the reduced pyran moiety, since the Δ1,8a double bond, as usually present in azaphilonoid compounds, is responsible for the high reactivity toward primary amines, and thus toxicity [1, 3–5]. Though the reactivity of acetosellin (1) with regard to amines has not been studied to the best of our knowledge, it can be assumed that acetosellin (1) is not as reactive under physiological conditions as other azaphilones. This study targets azaphilonoid metabolites in Epicoccum nigrum strain 749. Besides acetosellin (1), the novel metabolite 5′,6′-dihydroxyacetosellin (2) was obtained. Biosynthetically,

*

Dedicated to Prof. Max Wichtl in recognition of his outstanding contribution to pharmacognosy research.

Hufendiek P et al. Biosynthetic Studies on …

Planta Med

Results and Discussion E. nigrum strain 749 was isolated from an unspecified green alga. Sequencing of its ITS-region as well as morphological characteristics revealed its identity (Fig. S1, Supporting Information). Cultivation experiments showed that the metabolite pattern of E. nigrum is strongly dependent on the conditions applied. In experiments on different cultivation media containing trace metal supplementation, acetosellin (1) was only found as a very minor product in E. nigrum. However, a major change in the metabolic pattern was observed by omitting the trace metal supplementation and applying constant incubation under artificial light. Using these conditions the fungus produced acetosellin (1) as the major metabolite and, additionally, the new derivative 2, i.e., 5′,6′-dihydroxyacetosellin. With the higher acetosellin yields, biosynthetic studies became feasible. In order to interpret the results from labeling studies, NMR spectral data for 1 needed to be completely assigned. Comparison

Downloaded by: University of Colorado. Copyrighted material.

Affiliations

of our NMR data with the 1D‑NMR resonances given by Nasini et al. [6] proved to be cumbersome. On the one hand, spectroscopic data were reported without direct assignment to the respective atoms [6], and on the other hand, deviations of 13C‑NMR values for supposedly C-10 (Δδ = 3.3), C-13 (Δδ = 9.5), C-14 (Δδ = 3.3), and C-1′ (Δδ = 5.8) were higher than expected. At this point we concluded that we either obtained a compound different than acetosellin (1), or the NMR data in [6] were not correct. Consequently, acetosellin (1) was re-isolated in sufficient amounts and its structure firmly deduced and confirmed from extensive NMR measurements using 1H-/13C‑NMR, DEPT-135, as well as 2D‑NMR data (▶ Fig. 2; Tables S1a – S2 and Figs. S2 – S8, Supporting Information). Confirmation of the absolute configuration was achieved by optical rotation and CD measurements, and comparison with the reported data. During the isolation of 1, a slightly more hydrophilic compound, i.e., compound 2, was obtained. The molecular mass of 428 Da (Δ = 34 Da in comparison to 1) already led to the assumption that compound 2 might be a dihydroxylated derivative of acetosellin (1). The exact mass confirmed the molecular formula to be C23H24O8, while the hypsochromic shift in the UV spectrum (Δ = 25 nm; Fig. S17, Supporting Information) pointed to a shorter conjugated polyene system. Indeed, 1H- and 13C‑NMR spectra (▶ Table 1; Figs. S11 – S16, Supporting Information) showed that the alkene signals of C-5′/H-5′ as well as C-6′/H-6′ were missing when compared to the spectra of acetosellin (1). In turn, two resonances indicative of hydroxymethine functions appeared at δ 3.75/4.06 (1H‑NMR) and δ 71.7/77.4 (13C‑NMR). Through HMBC and 1H-1H-COSY correlations (▶ Fig. 3), the 1′ to 7′ side chain was proven to be hydroxylated at C-5′ and C-6′, leading to a vicinal diol partial structure. Comparison of the NMR chemical shifts of all C and H atoms confirmed the core structure to be identical with that of acetosellin (1). The size of the coupling constants JH‑1′,H‑2′ = JH‑3′,H‑4′ = 15.2 Hz established the geometry of the double bonds Δ1′,2′ and Δ3′,4′ to both be E. However, the absolute configuration of C-5′and C-6′ could not be resolved during this study. The configuration at C-8 and C-3 is suggested to be identical to that of acetosellin (1), since the CD spectra of 1 and 2 are comparable (Figs. S9 and S19, Supporting Information) and compound 2 most probably is derived from the latter. The biosynthesis of acetosellin-type compounds is of interest, since they seem not to be produced from one polyketide chain, as it is the case for the majority of polyketides. An earlier attempt by Plitzko et al. [7] failed due to cessation of acetosellin (1) production by their fungal strain. During this study, we successfully incorporated [1-13C]acetate into 1, and could deduce a clear pathway to this metabolite. In order to achieve efficient incorporation of [1-13C]acetate, experiments on acetosellin (1) production were performed. As outlined above, the production of acetosellin (1) is heavily influenced by the medium composition and the illumination of the fungal culture (▶ Fig. 4 a, b). Strain 749, grown on Czapek-Dox agar medium under continuous illumination (CDA), proved to be an excellent producer of acetosellin (1) (42 mg/L culture medium). The addition of a trace metal solution (consisting of ZnSO4 and CuSO4) led to the downregulation of acetosellin production. Instead, epipyrones [8] appear as the main metabolites (112 mg/

▶ Fig. 1 Azaphilonoid scaffold (modified from [1]) in comparison with acetosellin (1) and the related epicocconone (3). The characteristic core structure is shown in bold for compounds 1 and 3.

▶ Fig. 2 Structure of acetosellin (1) showing 1H-1H-COSY and key HMBC correlations.

L culture medium, fermentation in darkness). The latter are unusual pyronepolyenes with a C-galactosyl moiety, which is attached to the 2-pyrone partial structure. They reportedly convert spontaneously in solution at the C-glycosyl group to form three isomers. Due to experimental difficulties, isotope labeling using solid cultures has been performed in rare cases only. Instead, mostly liquid media have been used, in which precursors are usually much better incorporated into the respective target molecule. The E. nigrum strain 749, however, does effectively produce the natural products of interest exclusively on solid media. During this project, we thus applied a methodology using solid agar-based media, recently developed in our laboratory, where [1-13C]-


Planta Med


Original Paper

▶ Table 1 NMR data of 5′,6′-dihydroxyacetosellin (2) measured in MeOH-d4 at 600 MHz (1H‑NMR) and 150 MHz (13C‑NMR), respectively. Position

δ 13C, mult.a,b

δ 1H (J in Hz)b

1

65.5, CH2

a: 4.40 (1H, dt, 17.1; 3.3)

3

75.5, CH

3.72 (1H, m) †

4

33.3, CH2

a: 2.72 (1H, br d, 17.6)

b: 4.78 (1H, dd, 2.4; 17.1)

b: 2.90 (1H, m) 146.6, qC

6

118.5, qC

7

157.5, qC

8

86.6, qC

9

197.4, qC

10

129.0, qC

11

168.3, qC

12

110.6, qC

13

157.8, qC

14

117.8, CH

15

146.9, qC

▶ Fig. 3 Structure of 5′,6′-dihydroxyacetosellin (2) showing 1H-1HCOSY and key HMBC correlations.

6.93 (1H, s)

16

28.4, CH3

1.88 (1H, s)

17

65.4, CH2

3.75 (2H, m) †

1′

131.0, CH

7.18 (1H, d, 15.2)

2′

136.5, CH

6.75 (1H, dd, 10.2; 15.2)

3′

132.5, CH

6.62 (1H, dd, 10.2, 15.2)

4′

138.0, CH

6.08 (1H, dd, 6.5, 15.2)

5′

77.4, CH

4.06 (1H, m)

6′

71.7, CH

3.75 (2H, m) †

7′

18.9, CH3

1.22 (3H, d, 6.4)

Data referenced with Mestrenova using MeOH-d4 signal. a Multiplicities deduced from DEPT-135 experiments; b measured in MeOH-d4; † overlapped

labeled sodium acetate solution is added periodically onto the top of the growing fungal culture [9]. A medium scale (0.5 L) culture of E. nigrum strain 749 with added [1-13C]acetate was performed, and the 13C-labeled compound 1 was isolated. The substance was analyzed via 1H‑NMR, 13 C‑NMR, and LC‑MS measurements (Table S2 and Figs. S8, S10, and S20, Supporting Information). By comparing the NMR data of acetosellin (1) and the 13C-labeled compound, incorporation of eleven [1-13C]acetate units could be deduced (▶ Fig. 5). The following biosynthetic pathway can be envisioned: The biosynthesis of the pentaketide unit, presumably performed by a non-reducing PKS (NR‑PKS), starts from C-17 through to C-1 to form rings B and C (▶ Fig. 6 II). Aldol condensation between C-10 and C-5 leads to the formation of ring B. During the synthesis, two hydroxylations have to take place to form the hydroxy groups at C8 and C-17, since their respective position reveals that they most likely do not originate from acetate. As a consequence of the oxygenation at C-8, a dearomatization of the 2-methyl orcinaldehyde partial structure is achieved. Subsequently, the azaphilone-typical


Planta Med

pyran quinone partial structure can be built via acetal formation and subsequent reduction of ring C. The methyl group at C-8, i.e., CH3-16, presumably arises from methylation via S-adenosyl methionine (SAM) already during chain elongation. This notion is supported by the fact that methylation of polyketides by methyltransferases takes place at the methylene carbon in the α-position to the thioester of the growing chain [10]. In the case of acetosellin (1), C-8 does not show 13 C-labeling and consequently must have originated from a methylene group during chain elongation. C-16 is neither part of the polyketide backbone nor does it show 13C-labeling. Alternatively, the methyl group at C-8 could also be formed through incorporation of a methylmalonyl extender unit. However, to the best of our knowledge, there is no precedence for this in the literature on the biosynthesis of similar azaphilones. Also, the same methylation pattern and the responsible methyltransferases have been reported as part of the biosynthetic gene clusters of citrinin and azanigerone and the azaphilonoid Monascus pigments [1, 11, 12]. The biosynthesis of the hexaketide unit starting from C-7′ through to C-11 is presumably achieved by the action of a highly reducing PKS (HR‑PKS) (▶ Fig. 6 I). The interconnection with the pentaketide, i.e., C-17 to C-1, leads to the formation of rings A and D (▶ Fig. 6 III). During synthesis, the keto groups at C-6′, C4′ and C-2′, originally present in the nascent polyketide chain, are reduced to form a conjugated triene with double bonds at C6′, C-4′, and C-2′. The carbonyl group at C-13 is enolized. Both polyketide units are suggested to be connected at two positions. On one side, the hydroxy group at C-8 and the carboxylic group at C-11 undergo an esterification. On the other side, a proposed Knoevenagel condensation between C-12 and the keto group at C-7 forms the γ-lactone ring D, possibly yielding epicocconone (3) as an intermediate. Compound 3 was isolated before from E. nigrum [13]. Another aldol condensation under reductive influence of, e.g., oxidoreductases, between C-6 and the keto group at C-15 is the most likely origin of the aromatic ring A, leading to the final product, acetosellin (1). During this study, antimicrobial activity could not be attributed to acetosellin (1), though a panel of 30 microbes was tested (Table S3, Supporting Information). Neither yeasts nor gram-positive or gram-negative bacteria were inhibited. Additionally, no cy-


5

▶ Fig. 4 a Cultures of E. nigrum strain 749 growing either on Czapek-Dox agar under constant illumination with artificial white light (CDA) (left) or on Czapek-Dox agar supplemented with 20 µM CuSO4 and 35 µM ZnSO4 (CDT) (right). Pictures were taken after 14 days of cultivation. Cultures grown on CDA mainly produce acetosellin (1). Fermentation on CDT leads to pronounced production of epipyrones [8]. b Comparison of acetosellin (1) production by E. nigrum strain 749 under different cultivation conditions. Data were obtained by LC-HRMS and processed by MZmine 2.21 [33]. The extracted ion chromatogram for the mass of acetosellin (395.1489 [M + H]+) is shown in the positive mode. Acetosellin elutes at ca. 8.9 min. Continuous line: Cultivation on Czapek-Dox agar in darkness. Dotted line: The extracted ion chromatogram of acetosellin from the same strain grown on Czapek-Dox agar supplemented with 20 µM CuSO4 and 35 µM ZnSO4 in darkness.

totoxic effect was found in HEK cells at a concentration of 30 µM (Fig. S21, Supporting Information). Testing against different cysteine and serine proteases did not show an inhibitory activity at a concentration of 20 µM. Past investigations, however, showed that azaphilones have diverse biological activities, including antimicrobial, anti-inflammatory, anthelmintic, cytotoxic, and enzyme inhibitory effects [1, 2, 6, 14–19]. Among the azaphilones, the herein described acetosellin (1) and the new metabolite 2 show distinctive features, which are rarely found in other members of this class, i.e., a naphthopyran core structure (rings A, B, C), including a dihydropyran (ring C) moiety. To the best of our knowledge, this naphthopyran moiety is a unique trait of acetosellin and its derivative (2). The dihydropyran substructure can be found in a very limited number of analogues, e.g., bulgarialactone D (8) (▶ Fig. 7).

Our hypothesis on the biosynthesis of acetosellin is supported by a number of studies on the formation of different azaphilones. Similar findings regarding the labeling pattern, e.g., for ochrephilone (4) from Penicillium multicolor [7, 20], lend support to our conclusions considered herein (▶ Fig. 7). Compound 4, however, shares only structural features with 1 regarding the lower part of the molecule. Instead, structural resemblance is most obvious between acetosellin (1) and azaphilonoid pigments derived from Monascus spp., even though the latter do not have the distinctive aromatic moiety in their structure. In the case of monascorubrin (5), monascoflavin (6, syn. monascin), and PP‑V (7), the condensation of a hexaketide, forming the azaphilonoid core structure, with an acyclic pentaketide chain was proposed. Any one of them showed a comparable incorporation pattern to 1 after feeding the culture with labeled acetate as a precursor [1, 21–23]. A number of supportive studies on the genetic level were also published regarding azaphilone biosynthesis, e.g., in the case of azanigerones and the Monascus azaphilonoid pigments [11, 24, 25]. These findings suggest that biosynthetic gene clusters, harboring polyketide synthase (PKS) and fatty acid synthase (FAS) genes as well as tailoring enzymes, are responsible for the formation of azaphilonoid pigments derived from Monascus spp. and Penicillium marneffei [12, 26, 27]. In each case, two units, i.e., a polyketide and a fatty acid chain, were found to be involved in the biosynthesis. According to Balakrishnan et al. [12], the azaphilone core structure in the Monascus azaphilonoid pigments [e.g., monascorubrin (5) and monascoflavin (6); ▶ Fig. 7] is synthesized by a non-reducing fungal PKS with a reductive release domain (NR-fPKS‑R) in Monascus pilosus [25]. In P. marneffei, which also produces monascorubrin (5), amongst other azaphilonoids, knockdown mutants of the putative PKS gene (pks3) did not produce the azaphilonoid pigments anymore [27]. In another study [24], Balakrishnan et al. showed that a canonical FAS, called MpFAS2, is needed for the production of the acyclic side chain of Monascus azaphilone pigments in Monascus purpureus. Knockout mutants did not produce the typical pigments, but mainly synthesized monascusone A, which lacks the acyclic side chain. Interestingly, the data from Balakrishnan et al. obtained with M. purpureus [24] indicate that the FAS MpFAS2 may be dedicated solely to secondary metabolite synthesis, since the deletion mutant did not show an altered cellular fatty acid profile and content. In Aspergillus nidulans, a similar combined PKS‑FAS gene cluster was also analyzed. This is, however, responsible for the production of the carcinogenic xanthone derivative sterigmatocystin [28]. Brown et al. proved the FAS to be responsible for the supply of a hexanoyl starter unit during sterigmatocystin biosynthesis. Azanigerones are azaphilonoid compounds, which are usually composed of two polyketide chains [e.g., azanigerone B (9); ▶ Fig. 7]. The formation of the pyran ring was extensively studied in Aspergillus niger [11]. A putative NR‑PKS gene was found. It presumably synthesizes a poly-β-keto acid, which via aldol condensation gives rise to an aromatic ring (compare with ▶ Fig. 6 II). Biosynthesis of such aromatic polyketides is known to take place by NR‑PKS, involving a product template (PT) domain mechanism [29]. Dearomatization of the orcinaldehyde partial structure in azanigerone biosynthesis was shown in vitro to involve a monooxygenase (AzaH) for the formation of the azaphilonoid back-


Planta Med


Original Paper

Materials and Methods General procedures Optical rotation was measured on a Jasco DIP 140 polarimeter (1 dm, 1 cm3 cell) operating at wavelength λ = 589 nm correHufendiek P et al. Biosynthetic Studies on …

Planta Med

▶ Fig. 5 Structure of acetosellin (1) with 13C-labeling pattern after the addition of [1-13C]acetate to the culture medium. Bold dots indicate labeled 13C atoms, i.e., C-1 of acetate. The dashed red line marks the two different polyketide units, from which product 1 originates.

sponding to the sodium D line at room temperature. ECD (electronic circular dichroism) spectra were taken on a Jasco J-810 CD spectropolarimeter. UV spectra were obtained using a Perkin Elmer Lambda 40 UV/Vis spectrometer. IR spectra were recorded as film using a Perkin-Elmer FT‑IR Spectrum BX spectrometer. All NMR spectra were recorded using either a Bruker Avance 300 DPX spectrometer operating at 300 MHz (1H) and 75 MHz (13C), and a Bruker Ascend 600 spectrometer operating at 600 MHz (1H) and 150 MHz (13C). Spectra were referenced using Mestrenova version 8.0.1–10 878 [32] (Mestrelab Research S. L.) to residual solvent signals with resonances at δH/C 2.04/29.8 (acetone-d6), δH/C 3.35/49.0 (methanol-d4), and δH/C 2.50/39.5 (DMSO-d6), respectively. High-resolution LC‑MS was performed on a micrOTOF‑Q mass spectrometer (Bruker) with an ESI source coupled to a HPLC Dionex Ultimate 3000 (Thermo Scientific). An EC10/2 Nucleoshell C18 column (250 × 4.6 mm, 2.7 µm, Macherey-Nagel) was used. The column temperature was set to 25 °C. MS data were acquired over a range from 100–3000 m/z in the positive mode. HPLC begins with 90 % H2O containing 0.1 % HOAc. The gradient starts after 1 min, ending up with 100 % acetonitrile (0.1 % acetic acid) in 20 min. Five µL of a 1-mg/mL sample solution was injected with a flow of 0.3 mL/min. Data were processed with MZmine 2.21 [33].

Isolation of the fungal sample During this study, the strain 749 of the species E. nigrum LINK (Pleosporales, Didymellaceae) was cultivated and studied. It was isolated from the surface of an unspecified green alga on water agar, collected in Cabrera (Spain). A specimen is stored at the Institute of Pharmaceutical Biology, University of Bonn. For the isolation of fungal strains, algal samples were pressed onto sterile agar plates (without sterilization) containing water agar (WA: 15 g agar, 1 L artificial sea water) or chitin-malt extract agar (CMA: 15 g agar, 2 g chitin from crab shells, 10 g malt extract, 1 L artificial sea water). The isolation media were supplemented with 250 mg/L benzylpenicillin and streptomycin sulfate, respectively. In the second step, algal samples were sterilized with


bone. AzaH was needed for the conversion of the aromatic intermediate into the typical pyran quinone ring system of azanigerone E, the bicyclic precursor, which still lacks a side chain. AzaH is thus proposed to dearomatize the orcinaldehyde moiety through a hydroxylation reaction. Subsequently, the connection of both precursor units, i.e., azanigerone E and the acyl side chain, likely occurs under the influence of the acyltransferase AzaD [25]. For both enzymes, AzaH and AzaD, correlating genes have been found in the Monascus azaphilonoid pigment gene cluster [25]. In the case of acetosellin (1), a biosynthetic gene cluster would most probably not encompass a FAS, as in the case of the Monascus pigments, but instead, two different PKS genes, as was found for azanigerones [11]. This is required, since the functionalities, i.e., double bonds and β-/δ-keto groups of the hexaketide portion (▶ Fig. 6 I), cannot be introduced by a canonical FAS [10]. Instead, the programming of a PKS would allow for the generation of the observed substitution pattern. In the case of acetosellin, a highly reducing polyketide synthase (HR‑PKS) is suggested to synthesize the hexaketide chain, which forms the Western part of the molecule (▶ Fig. 6 I). The biosynthesis of the pentaketide unit (▶ Fig. 6 II) is suggested to occur, in principle, as outlined for the azanigerone biosynthesis [25]. In contrast to the azanigerones, however, the connection of both biosynthetic units does not only occur via ester bond formation, but involves a number of further reactions (▶ Fig. 6 III). The latter are, in part, similar to those of Monascus azaphilonoid pigment formation [12, 26, 27]. The last proposed reaction step in acetosellin (1) formation, the reduction of epicocconone (3) and subsequent cyclization to form the naphthopyran moiety, was proposed before [7], but no similar case could be found in the literature. Regarding the new metabolite 2, it is not known whether the oxidation of the side chain is induced enzymatically or spontaneously in or outside the cell. Oxidation reactions forming an epoxy group at the double bonds of polyenes are widespread in nature, specifically in carotenoids. Spontaneous (nonenzymatic) and enzymatic epoxidation were observed and are an important part of the antioxidant function of carotenoids in plants [30, 31]. The epoxide can be cleaved by a nucleophilic attack of a water molecule under acidic or basic conditions. Hydrolysis of the epoxide could also have happened during the isolation of 2, since water (with 0.1 % TFA) was used in the mobile phase. Nevertheless, in MS experiments, the characteristic fragments of 2 were observed in the crude extracts of strain 749 growing on CDA medium as well as in guttation droplets of this strain on the same substrate. Thus, we propose a biosynthetic origin for 2. Taken together, a biosynthetic pathway for acetosellin as shown in ▶ Fig. 6 is suggested. Based on these results, further studies can now elucidate the detailed enzymatic machinery, thus clarifying the exact biosynthetic sequence for these intriguing metabolites.

▶ Fig. 6 Proposed biosynthetic route to acetosellin (1) via epicocconone (3) following the condensation of two polyketide-derived intermediates (A and B, in sequences I and II). Sequence III shows the connection of A and B, as well as further tailoring reactions. This scheme is based on findings from azaphilonoid pigments derived from Monascus spp. and azanigerones [11, 25, 26, 41], combined with 13C-labeling results obtained during the present study. The sequence of steps is chosen arbitrarily. SAM = S-adenosyl methionine, NADPH = nicotinamide adenine dinucleotide phosphate (reduced form).

70 % ethanol, cut into pieces, and placed on culture plates containing the abovementioned media. The fungal colonies growing on agar plates were then transferred to new culture plates containing MES medium (20 g barley malt extract, 15 g agar, 1 L artificial sea water) in order to purify them. The identity of the strains was determined by ITS sequencing. For this purpose, the ITS4-primer, described by White et al. [34], was used. Specimens of the strain 749 are deposited at the Institute for Pharmaceutical Biology, University of Bonn.

Fermentation and extraction CDA (2.5 L; consisting of 35 g Difco Czapek-Dox broth, 16 g agar, 1000 mL demineralized water) was prepared as a substrate and

distributed on 25 petri dishes (145 × 20 mm) after autoclaving. Inoculation was carried out using three agar plugs, from precultures, per plate. Inoculated plates were sealed with parafilm and incubated at 26 °C under constant artificial illumination with white light. Cultures were harvested after 30 days, minced with a blender, extracted four times successively with EtOAc, and evaporated to dryness at 40 °C with a rotary evaporator, yielding 576 mg of crude extract.

Isolation of compounds 1 and 2 The abovementioned crude extract of strain 749 was first defatted using liquid-liquid extraction. For this purpose, the sample was dissolved in 100 mL MeOH/water (80 : 20) and successively shaken


Planta Med


Original Paper

three times with 100 mL petroleum ether in a separation funnel. The defatted fraction (321 mg) was subjected to flash chromatography using a Reveleris X2 Flash Chromatography System with UV and ELS detectors. For this procedure, the sample was dissolved in MeOH and adsorbed onto 3.5 g C18 silica gel (Polygoprep 60 C18, 50 µm, Macherey-Nagel) using a solid loader. Separation was carried out using a Reveleris C18 flash cartridge (40 µm particle size, 12 g C18 silica). For the mobile phase, MeOH (solvent A) was used in combination with 0.1 % aqueous TFA (solvent B). A step gradient with a constant flow of 30 mL/min was applied starting at 20 % solvent A (t = 0–5 min), rising up to 100 % solvent A (t = 5– 17 min), and finishing the run isocratically with 100 % solvent A (t = 17–22 min). The separation process was observed using ELS and UV detection (220 nm, 254 nm, 340 nm), and fractions were collected automatically according to the detector signal. Acetosellin (1, 104 mg) was eluted after 14 min (80 % solvent A). In fraction 6, eluting at 10 min, an analogue of acetosellin was detected, i.e., 5′,6′-dihydroxyacetosellin (2, tR ≈ 10 min, 58 % solvent A, 6 mg).


Planta Med

Labeling experiments For the labeling experiment, strain 749 was cultivated for 3 weeks on 500 mL of CDA (35 g Difco Czapek-Dox broth, 16 g agar, 1000 mL demineralized water) under continuous illumination with artificial light at 26 °C. A solution was prepared containing 55.6 mg/mL sodium [1-13C, 99 %]acetate (Cambridge Isotope Laboratories, Inc.) in demineralized H2O. A volume of 3 mL of this solution was added to each Fernbach flask during the cultivation period on days 4, 7, and 10, respectively. The solution was spread evenly on the agar surface in order to enable every part of the mycelium to get in contact with the labeled precursor. The total amount of added precursor was 500 mg sodium[1-13C, 99 %]acetate per flask, which equals a concentration of 2 g/L (= 0.024 mol/ L) medium. After 21 days, the labeled cultures (medium and mycelium) were minced using a blender and extracted three times successively with EtOAc. The solvent was evaporated at 40 °C using a rotary evaporator, yielding 111 mg of crude extract. Acetosellin (1): Red-orange amorphous powder (104 mg, 20 42 mg/L). α20 D + 264 (c 0.1, MeOH) (Nasini et al. [6]): αD + 283 (c 0.1, MeOH)]. ECD (c 0.25 mg/mL, MeOH): λmax (Δ ε) 335 (− 8.1);


▶ Fig. 7 Acetosellin (1) from E. nigrum, and azaphilones isolated from Monascus spp. [monascorubrin (5), monascoflavin (6), PP‑V (7)], P. multicolor [ochrephilone (4)]. They show a high structural similarity and a similar pattern of acetate incorporation after labeling studies (indicated by bold bonds). From Bulgaria inquinans, an additional metabolite is shown with a partly hydrated pyran moiety, similar to acetosellin [bulgarialactone D (8)] [40]. Azanigerone B (9) is shown as an example for a group of azaphilones, i.e., azanigerones, found in A. niger [11].

Original Paper

Bioassays Cell viability assay Cell viability was assessed using a fluorimetric detection of resorufin (CellTiter-Blue Cell Viability Assay, Promega; Fig. S21, Supporting Information). HEK293 cells (from ATCC) were seeded at a density of 25 000 cells per well into black 96-well poly-D-lysinecoated plates with a clear bottom. Three hours after seeding, cells were treated with 0.3 % DMSO or the compound dissolved in medium for 24 h. To detect cell viability, CellTiter-Blue reagent was added and cells were incubated for 1 h at 37 °C according to the manufacturerʼs instructions. Fluorescence (excitation 560 nm, emission 590 nm) was measured using a FlexStation 3 Benchtop Multimode Plate Reader and data are expressed as percentage of cell viability relative to DMSO control. The cytotoxic drug etoposide was used as a positive control.

Antimicrobial assays

Acknowledgements HR‑LC-ESIMS measurements were partly recorded by Marc Sylvester (Institute for Biochemistry and Molecular Biology, University of Bonn). The authors thank Anna-Christina Schulz-Fincke for performing the protease assays. The project TUR_WTZ‑044 was funded by the Bundesministerium für Bildung und Forschung (BMBF), the Deutsche Zentrum für Infektionsforschung (DZIF), and the NRW graduate research school Biotech-Pharma.

Conflict of Interest The authors declare no conflict of interest.

References [1]

Gao JM, Yang SX, Qin JC. Azaphilones: chemistry and biology. Chem Rev 2013; 113: 4755–4811

[2]

Osmanova N, Schultze W, Ayoub N. Azaphilones: a class of fungal metabolites with diverse biological activities. Phytochem Rev 2010; 9: 315–342

[3]

Coghlan DR, Mackintosh JA, Karuso P. Mechanism of reversible fluorescent staining of protein with epicocconone. Org Lett 2005; 7: 2401– 2404

[4]

Wei WG, Yao ZJ. Synthesis studies toward chloroazaphilone and vinylogous γ-pyridones: two common natural product core structures. J Org Chem 2005; 70: 4585–4590

[5]

Lin TF, Yakushijin K, Büchi GH, Demain AL. Formation of water-soluble Monascus red pigments by biological and semi-synthetic processes. J Ind Microbiol 1992; 9: 173–179

[6]

Nasini G. Structure and absolute configuration of acetosellin, a new polyketide from a phytotoxic strain of Cercosporella acetosella. Tetrahedron Lett 2002; 43: 1665–1668

[7]

Plitzko I. Zur Biosynthese des Borrelidins sowie Isolierung und Strukturaufklärung von Sekundärmetaboliten aus marinen und terrestrischen Mikroorganismen. Phd thesis, University of Göttingen, 2007. Available at http://ediss.uni-goettingen.de/bitstream/handle/11858/00–17350000–0006-AC9D-3/plitzko.pdf?sequence=1. Accessed June 3, 2015

[8]

Van Ginkel R, Selwood A, Wilkins AL, Ford S, Calder C. Anti-microbial compositions. Patent US2012/0108526 A1; 2012

[9]

Nazir M, El Maddah F, Kehraus S, Egereva E, Piel J, Brachmann AO, König GM. Phenalenones: insight into the biosynthesis of polyketides from the marine alga-derived fungus Coniothyrium cereale. Org Biomol Chem 2015; 13: 8071–8079

Antimicrobial activity was assessed following the protocol published previously [35].

Enzyme assays The inhibitory activity towards different cysteine and serine proteases was followed spectrophotometrically by observing the cleavage of chromogenic substrates and detecting the formation of para-nitroaniline (pNA). For the human cysteine protease cathepsin L, the substrate Z‑Phe-Arg-pNA was used as described [36]. The serine proteases were assayed accordingly, human leukocyte elastase (HLE) with MeOSuc-Ala-Ala-Pro-Val-pNA [37], bovine trypsin with Suc-Ala-Ala-Pro-Arg-pNA [38], and bovine chymotrypsin with Suc-Ala-Ala-Pro-Phe-pNA [39], respectively. For each assay, a final concentration of 20 µM of compound 1 was tested in duplicate.

Supporting information Spectroscopic data including UV, CD, IR, 1H NMR, 13C NMR, 2D‑NMR, and LCESIMS, and data of bioassays of compounds 1 and 2 as well as labeling experiments are available as Supporting Information.

[10] Cox RJ. Polyketides, proteins and genes in fungi: programmed nanomachines begin to reveal their secrets. Org Biomol Chem 2007; 5: 2010 [11] Zabala AO, Xu W, Chooi YH, Tang Y. Characterization of a silent azaphilone gene cluster from Aspergillus niger ATCC 1015 reveals a hydroxylation-mediated pyran-ring formation. Chem Biol 2012; 19: 1049–1059 [12] Balakrishnan B, Karki S, Chiu SH, Kim HJ, Suh JW, Nam B, Yoon YM, Chen CC, Kwon HJ. Genetic localization and in vivo characterization of a Monascus azaphilone pigment biosynthetic gene cluster. Appl Microbiol Biotechnol 2013; 97: 6337–6345 [13] Bell PJL, Karuso P. Epicocconone, a novel fluorescent compound from the fungus Epicoccum nigrum. J Am Chem Soc 2003; 125: 9304–9305 [14] Akihisa T, Tokuda H, Ukiya M, Kiyota A, Yasukawa K, Sakamoto N, Kimura Y, Suzuki T, Takayasu J, Nishino H. Anti-tumor-initiating effects of monascin, an azaphilonoid pigment from the extract of Monascus pilosus fermented rice (red-mold rice). Chem Biodivers 2005; 2: 1305–1309


Planta Med


214 (+ 9.5) nm; 1H-/13C‑NMR data see Tables S1a and S1b, Supporting Information. HR‑ESI‑MS m/z 412.1762 [M + NH4]+; 395.1483 [M + H]+ (calcd. for C23H23O6+, 395.1489); 377.1385 [M – H2O + H]+ (calcd. for C23H21O5+, 377.1384). 5′,6′-Dihydroxyacetosellin (2): Yellow amorphous powder (6 mg, 2.4 mg/L). α20 D + 149 (c 0.5, MeOH). UV (CH3OH) λmax [nm] (ε [L · mol−1 · cm−1]) 206 (25 691), 315 (10 816) nm; ECD (c 0.3 mg/mL, MeOH): λmax (Δ ε) 332 (− 8.2), 216 (+ 5.2) nm; IR (νmax) 3357, 2914, 2850, 1738, 1694, 1614, 1453, 1261 cm−1. 1 H-/13C‑NMR data see ▶ Table 1; for UV and IR data, see Figs. S17 and S18, Supporting Information. LC‑ESI-HRMS m/z 411.1453 [M – H2O + H]+ (calcd. for C23H23O7+, 411.1438), 429.1553 [M + H]+ (calcd. for C23H25O8+, 429.1544), 446.1794 [M + NH4]+ (calcd. for C23H28NO8+, 446.1809).

[16] Stadler M, Anke H, Dekermendjian K, Reiss R, Sterner O, Witt R. Novel bioactive azaphilones from fruit bodies and mycelial cultures of the ascomycete Bulgaria inquinans (Fr.). Nat Prod Res 1995; 7: 7–14 [17] Wang GYS, Borgeson BM, Crews P. Pitholides A–D, polyketides from a marine tunicate-derived culture of Pithomyces sp. Tetrahedron Lett 1997; 38: 8449–8452 [18] Anke H, Kemmer T, Höfle G. Deflectins, new antimicrobial azaphilones from Aspergillus deflectus. J Antibiot (Tokyo) 1981; 34: 923–928 [19] Talontsi FM, Dittrich B, Schüffler A, Sun H, Laatsch H. Epicoccolides: antimicrobial and antifungal polyketides from an endophytic fungus Epicoccum sp. associated with Theobroma cacao. Eur J Org Chem 2013; 2013: 3174–3180 [20] Seto H, Tanabe M. Utilization of 13C‑13C coupling in structural and biosynthetic studies. III. Ochrephilone – a new fungal metabolite. Tetrahedron Lett 1974; 15: 651–654 [21] Ogihara J, Kato J, Oishi K, Fujimoto Y. Biosynthesis of PP‑V, a monascorubramine homologue, by Penicillium sp. AZ. J Biosci Bioeng 2000; 90: 678–680 [22] Kurono M, Nakanishi K, Shindo K, Tada M. Biosyntheses of monascorubrin and monascoflavin. Chem Pharm Bull (Tokyo) 1963; 11: 359–362 [23] Tam WT. Characterization of polyketide synthases in Penicillium marneffei. Phd thesis, University of Hong Kong, 2012. Available at http://hdl. handle.net/10722/197137. Accessed June 2, 2016 [24] Balakrishnan B, Kim HJ, Suh JW, Chen CC, Liu KH, Park SH, Kwon HJ. Monascus azaphilone pigment biosynthesis employs a dedicated fatty acid synthase for short chain fatty acyl moieties. J Korean Soc Appl Biol Chem 2014; 57: 191–196 [25] Bijinu B, Suh JW, Park SH, Kwon HJ. Delineating Monascus azaphilone pigment biosynthesis: oxidoreductive modifications determine the ring cyclization pattern in azaphilone biosynthesis. RSC Adv 2014; 4: 59405– 59408

[29] Crawford JM, Korman TP, Labonte JW, Vagstad AL, Hill EA, KamariBidkorpeh O, Tsai SC, Townsend CA. Structural basis for biosynthetic programming of fungal aromatic polyketide cyclization. Nature 2009; 461: 1139–1143 [30] Jahns P, Latowski D, Strzalka K. Mechanism and regulation of the violaxanthin cycle: the role of antenna proteins and membrane lipids. Biochim Biophys Acta 2009; 1787: 3–14 [31] Britton G. Structure and properties of carotenoids in relation to function. FASEB J 1995; 9: 1551–1558 [32] Cobas C, Seoane F, Vaz E, Bernstein MA, Dominguez S, Pérez M, Sýkora S. Automatic assignment of 1H‑NMR spectra of small molecules: A new robust and fast expert system based on fuzzy logic and probabilistic methods. Magn Reson Chem 2013; 51: 649–654 [33] Pluskal T, Castillo S, Villar-Briones A, Orešič M. MZmine 2: modular framework for processing, visualizing, and analyzing mass spectrometry-based molecular profile data. BMC Bioinformatics 2010; 11: 1 [34] White TJ, Bruns T, Lee S, Taylor JW. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. PCR Protoc Guide Methods Appl 1990; 18: 315–322 [35] Mehnaz S, Saleem RSZ, Yameen B, Pianet I, Schnakenburg G, Pietraszkiewicz H, Valeriote F, Josten M, Sahl HG, Franzblau SG, Gross H. Lahorenoic acids A–C, ortho-dialkyl-substituted aromatic acids from the biocontrol strain Pseudomonas aurantiaca PB‑St2. J Nat Prod 2013; 76: 135–141 [36] Frizler M, Lohr F, Lülsdorff M, Gütschow M. Facing the gem-dialkyl effect in enzymeinhibitor design: preparation of homocycloleucine-based azadipeptide nitriles. Chemistry 2011; 17: 11419–11423 [37] Gütschow M, Pietsch M, Themann A, Fahrig J, Schulze B. 2,4,5-Triphenylisothiazol-3(2-H)-one 1,1-dioxides as inhibitors of human leukocyte elastase. J Enzyme Inhib Med Chem 2005; 20: 341–347 [38] Sisay MT, Steinmetzer T, Stirnberg M, Maurer E, Hammami M, Bajorath J, Gütschow M. Identification of the first low-molecular-weight inhibitors of matriptase-2. J Med Chem 2010; 53: 5523–5535 [39] Sisay MT, Hautmann S, Mehner C, König GM, Bajorath J, Gütschow M. Inhibition of human leukocyte elastase by brunsvicamides A–C: cyanobacterial cyclic peptides. ChemMedChem 2009; 4: 1425–1429

[26] Yang Y, Liu B, Du X, Li P, Liang B, Cheng X, Du L, Huang D, Wang L, Wang S. Complete genome sequence and transcriptomics analyses reveal pigment biosynthesis and regulatory mechanisms in an industrial strain, Monascus purpureus YY‑1. Sci Rep 2015; 5: 8331

[40] Musso L, Dallavalle S, Merlini L, Bava A, Nasini G, Penco S, Giannini G, Giommarelli C, De Cesare A, Zuco V, Vesci L, Pisano C, Dal Piaz F, De Tommasi N, Zunino F. Natural and semisynthetic azaphilones as a new scaffold for Hsp90 inhibitors. Bioorg Med Chem 2010; 18: 6031–6043

[27] Woo PCY, Lam CW, Tam EW, Lee KC, Yung KK, Leung CK, Sze KH, Lau SK, Yuen KY. The biosynthetic pathway for a thousand-year-old natural food colorant and citrinin in Penicillium marneffei. Sci Rep 2014; 4: 6728

[41] Balakrishnan B, Chandran R, Park SH, Kwon HJ. Delineating citrinin biosynthesis: Ctn-ORF3 dioxygenase-mediated multi-step methyl oxidation precedes a reduction-mediated pyran ring cyclization. Bioorg Med Chem Lett 2016; 26: 392–396

[28] Brown DW, Adams TH, Keller NP. Aspergillus has distinct fatty acid synthases for primary and secondary metabolism. Proc Natl Acad Sci 1996; 93: 14873–14877


Planta Med


[15] Akihisa T, Tokuda H, Yasukawa K, Ukiya M, Kiyota A, Sakamoto N, Suzuki T, Tanabe N, Nishino H. Azaphilones, furanoisophthalides, and amino acids from the extracts of Monascus pilosus-fermented rice (Red-Mold Rice) and their chemopreventive effects. J Agric Food Chem 2005; 53: 562– 565